Sarah Botterman Prof. Marc Hooghe

The Relation between Religion and Civic Engagement. A Comparative Analysis of the Impact of Religious Competition in Belgium and the Netherlands (1996-2006). Sarah Botterman Prof. Marc Hooghe Department of Political Science, Katholieke Universiteit Leuven Email: Sarah.Botterman@soc.kuleuven.be; Marc.Hooghe@soc.kuleuven.be Paper presented at the Dutch-Belgian Political Science Conference Leuven, 27-28 May 2010

Abstract The influence of religion on the public sphere has been declining over the past decades. However, the question remains whether the relationship between religion and civic engagement has undergone similar changes over time. Starting from the secularization thesis and the religious competition theory, we postulate the following hypotheses: (1) the relation between church practice and civic engagement is positive, but weakens over time and (2) religious pluralistic cultures with competition between denominations cause a strong and positive relationship between church practice and civic engagement. We investigate these hypotheses for Catholics and Protestants in Belgium (without competition) and the Netherlands (with competition). The results give no support to the secularization hypothesis nor to the religious competition hypothesis. We do not find a declining effect between religion and civic engagement over the years. Moreover, the so-called lazy monopoly church in Belgium is more strongly related to civic engagement in secular voluntary associations than the Protestant denomination in the Netherlands. 1

Introduction In most studies on civic engagement and social capital, religious involvement still turns out to be an important predictor for engagement levels. Religious people are more active in voluntary associations, donate more money to charity goals, and on an aggregate level there is even a positive correlation between religiosity and the number of blood donators and the volume of blood donations (Bekkers and Veldhuizen 2008; Bekkers and Schuyt 2008). The positive association between religiosity and civic engagement is certainly not a new one. Already in the 19th century, de Tocqueville (1969[1835]) proclaimed that religion was an essential element in the maintenance of the young democracy of the United States. For de Tocqueville, it did not matter which religion you adhered, as long as you did: Society has nothing to fear or hope from another life; what is most important for it is not that all citizens should profess the true religion, but that they should profess religion (de Tocqueville, 1969[1835]: 290). This relation was repeatedly found in many recent studies as well (see for instance Wilson and Janoski 1995; Putnam 2000; Becker and Dhingra 2001; Lam 2002; Uslaner 2002; Lam 2006; Ruiter and de Graaf 2006). To summarize, Greeley stated: Religion is (at least potentially) a powerful and enduring source of social capital (Greeley 1997: 592-93). After all, religious people are learned by their faith to act more altruistic, and are subsequently encouraged to participate in religious networks. Indeed, they feel it as a moral obligation to help others, to volunteer and to work for charity (Wuthnow 1991; Bekkers and Schuyt 2008). Through these mobilization stimuli, religious people come into contact with other social networks as well, reinforcing the participation drive that can become a kind of social norm, postulated by the religious group. Schwadel voices it in this way: what is preached from the pulpit and talked about in the pews influences church members activities, not just in the church but also outside of the church (Schwadel 2005: 169). Being 2 member of a religious group can thus be similar to being member of a voluntary association, as both have similar outcomes, providing members with certain skills to attain social capital and to integrate in their community (Wilson and Janoski 1995; Ammerman 1997; Greeley 1997; Lam 2002; Kwak et al. 2004). Churches bring people together at a regular basis and thus create a stable framework for several types of social connectedness (Cassel 1999). Locally interwoven in their community, they play an important role in the creation of social capital. Contacts, created by church communities, are personal and thus have a potential higher impact on the general societal mobilization. In this way, a social entity is created between church members (Verba et al. 1995). 2

If religion, however, is really such an important source of civic engagement, this spells trouble for the future development of participation levels. Certainly the West European countries, in contrast to the United States, have evolved into highly secularized countries, as levels of religiosity have sharply been declining in previous decades. Looking at international data of the European Value Study or the International Social Survey Programme, we see a decline in most European countries with regard to religious belonging and religious behaving. For instance, while in the 1960s still a large majority of the population in the Netherlands (82 percent) indicated to belong to a denomination, this was down to 58 percent in 2008 (Centraal Bureau voor de Statistiek 2009; Knippenberg 1992). Church attendance rates dropped significantly as well. In Britain, for instance, in 1950 still 15 percent attended church weekly while this percentage dropped to being less than 10 percent at the end of the previous century. Only in countries such as Ireland and Poland, the percentage of churchgoers has been relatively stable and the decline minimal. The Nordic countries, on the other hand, are highly secularized. For instance, only 15 percent of the Swedish population stated they believed in God in 1990. However, although most West European countries experienced the process of declining religiosity, we do not expect to find identical relationships between religion and civic engagement for all countries, as we also take into account religious competition. While some West European countries have a historically predominant religion, others are religiously pluralistic. For example, in the third wave of the European Value Study of 1999/2000, 99.3 percent of the Italian respondents who indicated to belong to a denomination, were member of the Roman Catholic Church, compared to 0.4 percent who belonged to the Protestant Church and 0.1 percent to the Orthodox Church. On the other hand, in Latvia, 33.2 percent of the respondents who indicated to belong to a denomination, were member of the Roman Catholic Church, compared to 28.6 percent who belonged to the Protestant Church and 28.3 percent to the Orthodox Church (Halman 2001). In this paper, we conduct a double case study on Belgium and the Netherlands as the ideal comparable countries. Both countries can be considered as similar cases, with regard to education and income level, and level of secularization. They are different, however, on the theoretically crucial element of religious competition. In what follows, we first review current theories on secularization and religious competition, before we present the cases of Belgium and the Netherlands. Subsequently we develop our two hypotheses, and introduce data and 3

methods. In the discussion section, we investigate the relevance of our findings, both from the point of view of secularization theories as of theories on religious competition. Religion and civic engagement in the low countries Religion can be considered as a multi-dimensional concept (Stark and Glock 1968; Dobbelaere 1981). A simple division of the meaning of religion can be made by focussing on the function that it has for the individual or for the group. Regarding the individual component, we can study religious beliefs and ideas; regarding the group component, we can study religious rituals, practices and worship. In this article, we will focus on this latter, external function of religion, namely the social component of religion. We can see religion as a sub-community within the society, creating ties and social networks between its members and beyond, fostering integration in a society and creating links with the civil society. As Cipriani (1998) argues, there is no religion if there is no group or form of church community. In this way, churches can be seen as moral communities, fulfilling their social function for its members, just like secular voluntary associations do. This reasoning was already acknowledged by Durkheim as he defined religion as: un système solidaire de croyances et de pratiques relatives à des choses sacrées, c est-à-dire séparées, interdites, croyances et pratiques qui unissent en une même communauté morale, appelée Église, tous ceux qui y adhèrent (Durkheim 1912: 65). Secularization thesis We start with the secularization theory that states that over the past decades, the influence of religion has been declining and losing its grip on society. However, we have to be careful when using the term secularization, as it is a multilayered concept. Indeed, secularization can be seen at the societal level, the institutional level or at the individual level (Dobbelaere 1981). If we look at the individual level, it is moreover contested that secularization has declined the level of religious beliefs and ideas. In fact, one can speak of a religion à la carte, the idea of creating your individual interpretation of religion, that has been gaining ground. Sociologists speak in the case of declining church practice in terms of believing without belonging (Luckmann 1967; Davie 1994; 2000). Although the majority or at least a significant part of the population still beliefs and finds spirituality and transcendence 4

important, they do not longer translate this in memberships of organized religions, never mind by participating in church masses. This trend of a religion à la carte is a typical component of the trend of individualization. A strict interpretation states that individualized belief will be less intense and less participatory in nature. In this study, we focus on the effects of secularization on the institutional level. Does institutional secularization, the declining church rates impede the relationship between religiosity and civic engagement? Secularization has a large impact at the institutional level, as the number of churchgoers is rapidly declining (Halman and Rijs 1999). It is precisely this regular church practice that is said to be responsible for a stronger participation in the civil society. The individual believing without belonging may thus be as valuable as others, however, the individual option creates less systematically mobilization stimulations to interact within the society and thus less participation. Only the organized religions, with their ritual meetings can lead to a reinforcement of societal participation. Thus, looking at the processes of individualization and secularization, the diminishing influence of religion on the secular and civil society lets us expect that the positive impact of organized religion on voluntary participation will decline over time. We postulate the following hypothesis: Hypothesis 1. As a result of secularization, the relation between church practice and membership in voluntary associations weakens over time Religious competition thesis Next to the secularization theory, the religious competition theory is equally important (Chaves and Gorski 2001). The basic assumption of this approach is that the religious market is an open free competitive market. As many authors (such as Zaleski and Gorski 1995; Chaves and Gorski 2001; Voas et al. 2002) investigate this relation, searching for the optimal size of church congregations and trying to solve the free-rider problem, there is no consensus on the relationship between size, competition and effectiveness. Some authors find a positive (Finke and Stark 1988; 1989; Iannaccone 1991; Finke 1992) or negative (Breault 1989; Land et al. 1991; Olson 1999) relation between religious competition and religious participation, while others find no significant relation (Christiano 1987; Chaves and Cann 1992; Voas et al. 2002). Most authors (Finke and Stark 1992; Stark and Iannaccone 1994) argue that religious pluralistic cultures with a religious competition will generate higher church affiliations, 5

higher institutional religiosity and thus will create more efficient and effective congregations and denominations. On the contrary, religious monopolies are denoted as lazy and complacent churches that create less affiliation (Stark 1998). Since they do not have to face competitors, it is argued that historically dominant churches will lack vigor, and therefore will be characterized by lower levels of commitment and mobilization. However, there is a counter argument stating that religious pluralism will decrease religious beliefs and practices, as potential believers might become confused by the abundant choice in religious belief systems in their society. Indeed, as Berger (1967) argues, the more denominations and beliefs there are, the less plausible each of them seems. This effect of plausibility will thus decline the religious beliefs and practices. Also Froehle (1994) voices that there is need to have a critical mass to foster participation in general and a flourishing community life. He states that there is no need for competition, but there is need for tolerance and a large enough denomination to enable it to have a positive effect on community life and civil society. Still, Stark et al. are convinced of their religious competition statement as they argue: The theory is not absent today, nor is it about the United Stated it purports to be general (Stark et al. 1995: 436). It has to be noted that this theory was originally developed only on to explain intra-church participation. In this article, however, we want to extend this literature by looking at the relationship between religious competition and civic engagement in the secular civil society, being participation in voluntary associations. Following the reasoning of the most prominent authors in the field, we assume that a monopoly of a church in a country will create a lazy church, while competition will create more incentives for churches to get people actively involved in society. Religions facing competition from other denominations will behave in a more active manner than monopoly denominations that do not have to face any competition. We postulate therefore the following hypothesis: Hypothesis 2. Religious competition leads to a stronger effect of church practice on membership in voluntary associations. The main research question in this study is thus whether competition changes the effect of church practice on participation in voluntary associations. Do churches encourage participation in associational life or does it impede members, as they have to compete against other religions as well? 6

The Belgian and Dutch Case In this study we will compare Belgium and the Netherlands. This offers a good case study as it allows us to investigate two denominations (Catholic versus Protestant), and a context without (Belgium) and with (the Netherlands) religious competition. If we investigate the secularization process in Belgium and the Netherlands, we can see in Table 1 that there are substantial differences depending on denomination and context. Belgium is largely a monoreligious culture, characterized by a religious quasi-monopoly of the Roman Catholic Church. While the process of secularization has reduced the influence and power of the Roman Catholic Church, there is no viable competition of any other denomination. The Netherlands, on the other hand, is a religiously pluralistic country with the two strongest denominations being Roman Catholicism and Protestantism. The Netherlands thus differ strongly from Belgium, as there is a religious competition between the different denominations. We do not include other, smaller, denominations, as they would not play a relevant role in our analysis due to the low number of adherents. [insert Table 1 around here] When comparing the two cases, first, it has to be noted that no population data are available for the whole of Belgium. Therefore, we will limit the analysis to the Flemish, northern region in Belgium, with almost 60 percent of the total Belgian population. In 1996, 20 per ent of the population of the Flemish region indicated to have no religion, while in the Netherlands, already a majority of 61 percent indicated to have no religion. Over time these figures remained rather stable, in both societies. However, we can notice the difference in question wording in 1996 and 2006 in the Netherlands. In 1996, the denomination question was posed in the Netherlands as a one-step question: To which denomination do you belong? In 2006, it was a two-steps question: 1. Do you belong to a denomination? 2. To which denomination do you belong? It is expected that a two-steps question will yield higher estimates than a one-step question, but in this table, it seems not to be the case. If we compare the percentages to the ones in the European Value Study that uses a two-step question, we see a slightly smaller percentage of 54 percent in the third wave of 1999/2000. Looking at the percentages that indicated to belong to a denomination, a large majority of 78 percent in Belgium indicated to be Catholic and less than one percent indicates to be Protestant, while in 7

the Netherlands 18.5 percent indicated to be Catholic, in comparison to 11.7 percent that indicated to be Protestant. Further, we notice the evolution of church practice between 1996 and 2006 for all respondents and for the separate denominations that we study. Church practice is clearly declining during this observation period. While Catholics in both Flanders as the Netherlands seem to experience the effects of secularization quite severely between 1996 and 2006 (a decline of almost 10 percent for both groups), we see that Protestants in the Netherlands have not been experiencing such a strong decline in regular church practice. Data and Methods Data sources We use two representative cross-sectional population surveys, one for Flanders, the largest region in Belgium and one for the Netherlands. For the Flemish Region of Belgium, we use the annual face-to-face survey Social and Cultural Shifts, conducted by the Flemish regional government since 1996. This cross-sectional survey includes ca. 1,500 adult respondents every year. In the Netherlands, a similar face-to-face survey is conducted biennially by the Dutch Social and Cultural Planning Office, the Cultural Changes survey. The number of respondents in every wave is around 2,000. The pooled dataset of the Belgian Social and Cultural Shifts surveys between 1996 and 2006 (eleven waves) consists out of 17,714 respondents. Since both our hypotheses only refer to church practice, we selected only those respondents who indicated to be Catholic, leaving us with 12,956 respondents. 1 As mentioned before, we do not separately take the group of Protestants in Flanders, as this is a small group of less than 1 percent of the population, so no quantitative conclusions can be drawn. We use the Dutch Cultural Changes survey with only those respondents who indicated to belong to one of the two largest denominations, giving us two datasets consisting only of Catholics or Protestants. Since participation is only questioned in 1996 and 2006, we can only use these two waves, giving us a smaller number of research populations of 813 Catholics and 613 Protestants over the two years (434 Catholics in 1996, 396 Catholics in 2006; 374 Protestants in 1996, 250 Protestants in 2006). 8

Dependent variable The dependent variable in this study is being a member in a voluntary association. On the one hand we know whether or not the respondent is a member of a specific association; on the other hand we have a categorical variable summarizing the number of memberships. In the analysis, we limit ourselves to secular associations since including religious organizations would lead to a purely tautological relation. The dependent variable measures the civic engagement of the respondent. We look at the number of memberships. In the Belgian survey, we looked at the categories of voluntary associations that were questioned every year and coded the other associations in a residual category other associations. Ranging from 0 to 9 memberships in voluntary associations, we then categorised this dependent variable further into a categorical variable, with evenly distributed categories. The associations that were questioned over these 12 years were: (1) helping associations; (2) youth associations; (3) environmental associations; (4) cultural or arts associations; (5) socio-cultural associations; (6) sport associations; (7) neighbourhood associations; (8) union or work associations; (9) other associations. For the Dutch survey, we only have the 1996 and 2006 observation points, since voluntary membership was not included in the questionnaire in all other years. The range of memberships was again reduced to a categorical variable and was formed by the associations that were surveyed in both survey years, being (1) union associations; (2) hobby associations, including sport, cultural, arts and hobby associations; (3) political associations; (4) other associations. This last category was again a recoded variable of being member in an association that had been included in the questionnaire in just one of the two observations. Control variables The control variables were the same for Belgium as for the Netherlands. The determinants of individual voluntary association participation have been studied abundantly (Smith 1994; Rotolo 1999; Hooghe 2003). Therefore, we included control variables that have been shown to have a significant effect on participation levels. Looking at gender, classic research shows that men have higher overall membership rates in voluntary associations (Verba and Nie 1972; Curtis 1971). The relation with age tends to be curvilinear (Verba and Nie 1972; Curtis 1971). Age is recoded into three categories: (1) younger than 35 year; (2) between 35 and 55 9

years; and (3) older than 55 years. Next, higher educated persons have higher participation rates in voluntary associations (Verba et al. 1995). Education is recoded into three categories ranging from low education level (till lower secondary education), median education level (attained a higher secondary diploma) to high education (higher education). 2 Church practice is recoded into four categories, namely (1) never attends church; (2) rarely or only on religious Holy days; (3) monthly or more than once a month; (4) weekly or more than once a week. The Belgian survey data could be pooled from 1996 till 2007, so we also include year as a variable, ranging from 1 (survey in 1996) till 12 (survey in 2007). To look at the effect of year of survey on the relationship between church practice and participation, we further computed an interaction variable for the Belgian data. Both variables were standardised before multiplying, thus avoiding problems of multicollinearity (Jaccard et al. 1990). The Dutch survey data from 1996 and 2006 is pooled, thus we include year as a dummy variable, coding 1996 zero and 2006 one. Also in this case we computed an interaction variable following the same precautions. Methods We performed a logistic regression analysis, to ascertain whether participation is influenced by church practice. Subsequently, we included the interaction term in the models. Further, we looked at the differences in scope of participation, by categorising the dependent variable into no memberships, only one membership and two or more memberships and conducting multinomial regression analysis. As metrical variables caused too many empty cells, both dependent and independent variables were categorised. 10

Results First we investigate hypothesis 1. In Table 2, we conduct logistic regression analyses, predicting the likelihood of being a member of at least one voluntary association. [insert Table 2 around here] The analysis shows strong denominational differences. For Catholics (both in Belgium as in the Netherlands) church practice has a strong positive effect on voluntary participation. This effect is even stronger in Belgium, where regular churchgoers are 2.8 times as likely to be a member of an association as those who never go to church. Among Dutch Protestants, on the other hand, church practice is not significantly related to voluntary engagement. Looking at effects over time, again we see strong differences. There is a significant positive effect for Belgian Catholics (indicating more intensive membership over time), no effect for Dutch Catholics, and a significant negative effect for Dutch Protestants (indicating a decline over time). When including the interaction effect for the three groups (Model 2,4 and 6), we observe only a weakly significant but positive relation for Belgian Catholics (p =.054). Contrary to what secularization theory would predict, however, we do not observe a significant decline (and for Belgium even the opposite), so we can conclude that the relation between church practice and voluntary engagement certainly does not weaken over time. Our first hypothesis, therefore, does not receive confirmation. Looking at the other control variables, there are no surprising effects, with higher education levels for the highly educated and for men. Subsequently, we turn to the intensity of membership. In the following models, the number of voluntary memberships becomes the dependent variable (recoded in three categories). The models prove to be strongest for the Catholics, whether this is in Belgium or in the Netherlands (Tables 3 and 4). Among Catholics, church practice is clearly related to multiple memberships. Regular churchgoers tend more often to be a member of two or more voluntary associations. [insert Table 3 around here] [insert Table 4 around here] [insert Table 5 around here] 11

Looking at the intensity of participation among Dutch Protestants (Table 5), we can observe that church practice does not have a significant effect on multiple memberships. During the observation period, multiple membership even tends to decrease quite sharply among this group. Discussion We started this article with two clear hypotheses: the relation between church practice and participation should become weaker over time, and the relation should be strongest in a context of religious competition. Both hypotheses, however, did not receive confirmation. The results indicate, therefore, that context matters. For the secularization thesis, we can observe strong inter-denominational differences. While among Protestants the relation between church practice and participation indeed tends to become weaker, this is not confirmed for Catholics. Furthermore, religious competition does not seem to be important in keeping a denomination vigorous, as the strongest effects on participation were found in Belgium, a country without religious competition. Even when just comparing Catholics in Belgium and the Netherlands, it is clear that the lack of competition in Belgium resulted in more active believers. Relating to the secularization hypothesis, we even find a positive significant interaction effect of church practice and year of the survey, indicating that the effect of church practice becomes stronger over the years, at least in Belgium. Despite the fact that Catholics in Belgium do not encounter any competitors, they can hardly be called a lazy church, and they even become less lazy over time. What we do see, however, is that when studying the effect of religion on civic engagement, it is important to take into account various context variables. First, denomination clearly seems to matter, as we documented different effects, even for Catholics and Protestants in the same country. Second, country effects too should be taken into account, but contrary to expectations, it seems that denominations that do not encounter competition are more successful in mobilizing their members for various voluntary associations. 12

References Ammerman, Nancy T. 1997. Organized religion in a voluntaristic society. Sociology of Religion 58(3): 203-15. Becker, Penny E. and Pawan H. Dinghra. 2001. Religious involvement and volunteering: Implications for civil society. Sociology of Religion 62(3): 315-35. Bekkers, René and Theo Schuyt. 2008. And who is your neighbor? Explaining denominational differences in charitable giving and volunteering in the Netherlands. Review of Religious Research 50(1): 74-96. Bekkers, René and Ingrid Veldhuizen. 2008. Geographical differences in blood donation and philanthropy in the Netherlands: what role for social capital? Journal of Economic, and Social Geography 99(4): 483-96. Berger, Peter. 1967. The sacred canopy. New York: Doubleday. Breault, Kevin D. 1989. New evidence on religious pluralism, urbanism, and religious participation. American Sociological Review 54(6): 1048-53. Cassel, Carol A. 1999. Voluntary associations, churches, and social participation. Theories of turnout. Social Science Quarterly 80(3): 504-17. Centraal Bureau voor de Statistiek. 2009. Religie aan het begin van de 21ste eeuw. Den Haag: Centraal Bureau voor de Statistiek. Chaves, Marc and David E. Cann. 1992. Regulation, pluralism, and religious market structure. Rationality and Society 4(3): 272-90. Chaves, Marc and Philip S. Gorski. 2001. Religious pluralism and religious participation. Annual Review of Sociology 27: 261-81. Christiano, Kevin J. 1987. Religious diversity and social change: American cities, 1890-1906. New York: Cambridge University Press. Cipriani, Roberto. 1998. Sociology and religion: Durkheim and Simmel. In Secularisation and social integration. Papers in honor of Karel Dobbelaere, edited by Rudi Laermans, 37-44. Leuven: Leuven University Press. Curtis, James E. 1971. Voluntary association joining: a cross-national comparative note. American Sociological Review 36: 872-80. Curtis, James E., Grabb, Edward G. and Douglas E. Baer. 1992. Voluntary association membership in 15 countries. A comparative analysis. American Sociological Review 57(2): 139-52. Davie, Grace. 1994. Religion in Britain since 1945. Oxford: Blackwell. 13

Davie, Grace. 2000. Religion in modern Europe. A memory mutates. Oxford: Oxford University Press. De Tocqueville, Alexis. 1969[1835]. Democracy in America, translated by George Lawrence. Garden City: Doubleday. Dobbelaere, Karel. 1981. Secularization: A multi-dimensional concept. Current Sociology 29(2): 1 213. Durkheim, Emile. 1912. Les formes élémentaires de la vie religieuse. Paris: Presses Uuniversitaires de France. Finke, Roger. 1992. An unsecular America. In Religion and modernization: Sociologists and historians debate the secularization thesis, edited by Steve Bruce, 145-69. Oxford: Clarendon. Finke, Roger and Rodney Stark. 1988. Religious economies and sacred canopies: Religious mobilization in American cities. American Sociological Review 53(1): 41-49. Finke, Roger and Rodney Stark. 1989. Evaluating the evidence: Religious economies and sacred canopies. American Sociological Review 54(6): 1054-56. Finke, Roger and Rodney Stark. 1992. The churching of America, 1776 1990: Winners and losers in our religious economy. New Brunswick: Rutgers University Press. Froehle, Bryan T. 1994. Religious competition, community building, and democracy in Latin America: Grassroots religious organizations in Venezuela. Sociology of Religion 55(2): 145-62. Halman, Loek. 2001. The European Values Study: A third wave. Source book of the 1999/2000 European Values Study surveys. Tilburg: European Values Study. Halman, Loek and Ole Riis. 1999. Religion in secularizing society: the European's religion at the end of the 20th century. Tilburg: Tilburg University Press. Greeley, Andrew. 1997. Coleman revisited: Religious structures as a source of social capital. American Behavioral Scientist 40: 587-94. Hooghe, Marc. 2003. Why should we be bowling alone? Cross-sectional results from a Belgian survey on civic participation. Voluntas 14(1): 41-59. Iannaccone, Laurence R. 1991. The consequences of religious market structure. Rationality and Society 3: 156-77. Jaccard, James, Turrisi, Robert and Choi K. Wan. 1990. Interaction effects in multiple regression. Newbury Park: Sage. 14

Knippenberg, Hans. 1992. De religieuze kaart van Nederland: Omvang en geografische spreiding van de godsdienstige gezindten vanaf de Reformatie tot heden. Assen: Van Gorcum. Kwak, Nojin, Shah, Dhavan V. and R. Lance, Holbert. 2004. Connecting, trusting, and participating: The direct and interactive effects of social associations. Political Research Quarterly 57(4): 643-52. Lam, Pui-Yan. 2002. As the flocks gather: How religion affects voluntary association participation. Journal for the Scientific Study of Religion 41(3): 405-22. Lam, Pui-Yan. 2006. Religion and civic culture: A cross-national study of voluntary association membership. Journal for the Scientific Study of Religion 45(2): 177-93. Land, Kenneth C., Deane, Glenn and Judith R. Blau. 1991. Religious pluralism and church membership: A spatial diffusion model. American Sociological Review 56(2): 237-49. Luckmann, Thomas. 1967. The invisible religion. The problem of religion in modern society. Londen: Collier-MacMillan. Olson, Daniel V. A. 1999. Religious pluralism and U.S. church membership: A reassessment. Sociology of Religion 60(2): 149-74. Putnam, Robert. 2000. Bowling alone. The collapse and revival of American community. New York: Simon and Schuster. Rotolo, Thomas. 1999. Trends in voluntary association participation. Nonprofit and Voluntary Sector Quarterly 28(2): 199-212. Ruiter, Stijn and Nan D. De Graaf. 2006. National context, religiosity and volunteering: Results from 53 countries. American Sociological Review 71(2): 191-210. Schwadel, Philip. 2005. Individual, congregational, and denominational effects on church members civic participation. Journal for the Scientific Study of Religion 44(2): 159-71. Smidt, Corwin. 1999. Religion and civic engagement: A comparative analysis. Annals of the American Academy of Political and Social Science 565(1): 176-92. Smith, David H. 1994. Determinants of voluntary association participation and volunteering. A literature review. Nonprofit and Voluntary Sector Quarterly 23(3): 243-63. Stark, Rodney. 1998. Catholic contexts: Competition, commitment and innovation. Review of Religious Research 39(3): 197-208. Stark, Rodney, Finke, Roger and Laurence R. Iannaccone. 1995. Pluralism and piety: England and Wales, 1851. Journal for the Scientific Study of Religion 34(4): 431-44. Stark, Rodney and Charles Y. Glock. 1968. Patterns of religious commitment. Berkeley: University of California Press. 15

Stark, Rodney and Laurence R. Iannaccone. 1994. A supply-side reinterpretation of the secularization of Europe. Journal for the Scientific Study of Religion 33: 230-52. Uslaner, Eric M. 2002. Religion and civic engagement in Canada and the United States. Journal for the Scientific Study of Religion 41(2): 239-54. Verba, Sidney and Norman Nie. 1972. Participation in America: Political democracy and social equality. New York: Harper and Row. Verba, Sidney, Schlozman, Kay L. and Henry Brady. 1995. Voice and equality. Civic voluntarism in American politics. Cambridge: Harvard University Press. Voas, David, Crockett, Alasdair and Daniel V. A. Olson. 2002. Religious pluralism and participation: Why previous research is wrong. American Sociological Review 67(2): 212-30. Wilson, John and Thomas Janoski. 1995. The contribution of religion to volunteer work. Sociology of Religion 56(2): 137-52. Wuthnow, Robert. 1991. Between states and markets: The voluntary sector in comparative perspective. Princeton: Princeton University Press. 16

Table 1. Religious denomination and church attendance, by region (percentages) Flanders (Belgium) The Netherlands Denomination 1996 2006 1996 2006 Catholic 78.1 71.9 18.0 18.5 Protestant 0.3 0.2 15.5 11.7 Other religion 1.5 3.9 5.4 7.1 No religion 19.5 24.2 60.9 60.7 Church practice 1996 2006 1996 2006 Weekly 15.2 7.5 14.5 11.8 Periodically 13.0 8.6 10.2 10.1 Rarely 38.7 38.6 18.4 21.4 Never 33.2 45.3 56.8 56.6 Church practice Catholics 1996 2006 1996 2006 Weekly 18.6 8.9 23.1 15.4 Periodically 16.2 11.4 25.2 27.6 Rarely 42.6 47.0 36.0 38.1 Never 22.6 32.8 15.7 18.9 Church practice Protestants 1996 2006 1996 2006 Weekly 42.4 42.8 Periodically 25.5 23.2 Rarely 14.5 16.8 Never 17.7 16.8 Source: Social Cultural Change Studies (Belgium) and Cultural Changes (Netherlands) 17

Table 2. Logistic regression models: Effects of church practice on memberships in secular voluntary associations. Belgian Catholics (1996-2006) Dutch Catholics (1996/2006) Dutch Protestants (1996/2006) Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Independent Variables Coeff. Odds Coeff. Odds Coeff. Odds Coeff. Odds Coeff. Odds Coeff. Odds Constant -1.030*** 0.357-1.024*** 0.359 0.714* 2.043 1.124 3.077 0.996** 2.707 1.212* 3.361 (0.079) Year 0.181*** (0.006) Education (ref. = Low) High 0.668*** (0.058) Median 0.271*** (0.051) Church Practice (ref. = Never) Weekly + 1.016*** (0.074) Monthly + 0.616*** (0.071) Seldom 0.259*** (0.049) Age (ref. = < 35 year) 55 year -0.428*** (0.073) 35-55 year 0.020 (0.057) Gender (ref. = Woman) Man 0.424*** (0.042) (0.079) 1.198 0.180*** (0.006) 1.951 0.669*** (0.058) 1.311 0.272*** (0.051) 2.761 0.976*** (0.077) 1.852 0.600*** (0.072) 1.295 0.259*** (0.049) 0.652-0.423*** (0.073) 1.020 0.018 (0.057) (0.321) 1.197-0.254 (0.157) 1.951 0.248 (0.330) 1.312 0.488* (0.230) 2.653 0.671* (0.256) 1.822 0.531** (0.235) 1.295 0.573* (0.220) 0.655-0.719* (0.283) 1.019-0.300 (0.284) (0.588) 0.776-0.583 (0.426) 1.281 0.253 (0.330) 1.629 0.483* (0.230) 1.955 0.460 (0.359) 1.701 0.408 (0.278) 1.774 0.511* (0.232) 0.487-0.708* (0.283) 0.741-0.295 (0.284) (0.317) 0.558-0.522** (0.189) 1.288 0.101 (0.331) 1.620 0.160 (0.239) 1.585 0.160 (0.264) 1.503 0.003 (0.297) 1.668-0.178 (0.322) 0.493-0.326 (0.251) 0.744 0.411 (0.265) (0.614) 0.593-0.668 (0.403) 1.106 0.091 (0.332) 1.174 0.162 (0.239) 1.174 0.047 (0.382) 1.003-0.073 (0.350) 0.837-0.218 (0.336) 0.722-0.324 (0.251) 1.509 0.416 (0.265) 1.528 0.426 1.531 0.282 1.325 0.287 1.332 0.386* 1.471 0.386* 1.471 (0.042) (0.163) (0.163) (0.193) (0.193) Interaction 0.043 1.044-0.129 0.879-0.069 0.934 (0.022) (0.155) (0.166) Pseudo R² (Nagelkerke) 0.15 0.15 0.06 0.06 0.07 0.07 N 11,939 11,939 813 813 613 613 Note: Numbers in parentheses are standard errors. Dependent variables are dummies with active member in any kind of secular voluntary association coded one. p<0.10; *p<0.05; **p<0.01; ***p<0.001. 0.513 1.095 1.176 1.048 0.929 0.804 0.723 1.515 18

Table 3. Multinomial regression models: Effects of church practice on membership intensity in secular voluntary associations (Belgian Catholics). No memberships versus one membership No membership versus two or more memberships One membership versus two or more memberships Independent Variables Coeff. Odds Coeff. Odds Coeff. Odds Intercept -0.976*** (0.087) -2.670*** (0.098) -1.694*** (0.094) Year 0.132*** 1.141 0.239*** 1.270 0.108*** 1.114 (0.007) (0.008) (0.007) Education (ref. = Low) High 0.300*** 1.349 1.064*** 2.898 0.764*** 2.147 (0.065) (0.066) (0.060) Median 0.121* 1.129 0.466*** 1.594 0.345*** 1.412 (0.056) (0.060) (0.057) Church Practice (ref. = Never) Weekly + 0.781*** 2.183 1.319*** 3.738 0.538*** 1.713 (0.082) (0.087) (0.078) Monthly + 0.370*** 1.447 0.912*** 2.489 0.542*** 1.720 (0.080) (0.082) (0.077) Seldom 0.134* 1.143 0.421*** 1.524 0.287*** 1.333 (0.055) (0.059) (0.056) Age (ref. = < 35 year) 55 year -0.427*** 0.652-0.453*** 0.635-0.026 0.974 (0.080) (0.087) (0.082) 35-55 year -0.124 0.884 0.178** 1.195 0.302*** 1.352 (0.064) (0.067) (0.061) Gender (ref. = Woman) Man 0.270*** 1.310 0.611*** 1.842 0.341*** 1.407 (0.047) (0.049) (0.046) Pseudo R² (Nagelkerke) 0.17 N 11,939 Note: Numbers in parentheses are standard errors. p<0.10; *p<0.05; **p<0.01; ***p<0.001. 19

Table 4. Multinomial regression models: Effects of church practice on membership intensity in secular voluntary associations (Dutch Catholics). No memberships versus one membership No membership versus two or more memberships One membership versus two or more memberships Independent Variables Coeff. Odds Coeff. Odds Coeff. Odds Intercept 0.066 (0.358) -0.202 (0.377) -0.268 (0.358) Year 0.370* 1.448-0.888*** 0.412-1.258*** 0.284 (0.183) (0.186) (0.185) Education (ref. = Low) High -0.110 0.896 0.479 1.614 0.589 1.802 (0.400) (0.356) (0.354) Median 0.235 1.265 0.705** 2.024 0.470 1.600 (0.265) (0.255) (0.239) Church Practice (ref. = Never) Weekly + 0.258 1.295 1.088*** 2.968 0.830* 2.292 (0.302) (0.306) (0.322) Monthly + 0.159 1.173 0.927** 2.526 0.767* 2.154 (0.273) (0.284) (0.297) Seldom 0.528* 1.696 0.656* 1.927 0.127 1.136 (0.248) (0.273) (0.277) Age (ref. = < 35 year) 55 year -0.779* 0.459-0.644* 0.525 0.135 1.144 (0.313) (0.325) (0.294) 35-55 year -0.489 0.613-0.095 0.909 0.394 1.483 (0.315) (0.321) (0.285) Gender (ref. = Woman) Man 0.054 1.056 0.510** 1.665 0.455* 1.577 (0.189) (0.186) (0.186) Pseudo R² (Nagelkerke) 0.16 N 813 Note: Numbers in parentheses are standard errors. p<0.10; *p<0.05; **p<0.01; ***p<0.001. 20

Table 5. Multinomial regression models: Effects of church practice on membership intensity in secular voluntary associations (Dutch Protestants). No memberships versus one membership No membership versus two or more memberships One membership versus two or more memberships Independent Variables Coeff. Odds Coeff. Odds Coeff. Odds Intercept 0.495 (0.351) -0.013 (0.369) -0.508 (0.336) Year -0.091 0.913-1.030*** 0.357-0.939*** 0.391 (0.211) (0.224) (0.209) Education (ref. = Low) High -0.171 0.842 0.364 1.440 0.536 1.709 (0.379) (0.363) (0.331) Median -0.102 0.903 0.458 1.580 0.559* 1.750 (0.271) (0.268) (0.246) Church Practice (ref. = Never) Weekly + 0.186 1.205 0.124 1.132-0.062 0.940 (0.296) (0.303) (0.280) Monthly + -0.089 0.915 0.083 1.086 0.172 1.188 (0.336) (0.337) (0.315) Seldom -0.008 0.992-0.412 0.662-0.404 0.668 (0.357) (0.379) (0.356) Age (ref. = < 35 year) 55 year -0.589* 0.555 0.005 1.006 0.594* 1.812 (0.276) (0.298) (0.274) 35-55 year -0.050 0.951 0.944** 2.571 0.995*** 2.704 (0.291) (0.305) (0.266) Gender (ref. = Woman) Man 0.245 1.278 0.550* 1.733 0.304 1.356 (0.216) (0.219) (0.198) Pseudo R² (Nagelkerke) 0.12 N 613 Note: Numbers in parentheses are standard errors. p<0.10; *p<0.05; **p<0.01; ***p<0.001 21

Endnotes 1. We include those respondents who answered to the denomination question I am Catholic or I am Christian, but not Catholic. We include this latter answer category, as it was interpreted by respondents as a category that stands separate from the Catholic Church as institution. When, like in the Social and Cultural Shifts survey of 2008, this latter answer category is not offered, this percentage will indicate to be Catholic as well. 2. Family income could not be included as an indicator for socio-economic status because of the high number of missing cases. 22