Introduction

For almost a decade, the argument in favour of multiculturalism in Western Europe has been losing popularity (Vertovec and Wessendorf 2010). Whether we look in policy documents, in journalistic accounts or academic debates, very few seem to argue in favour of the cultural retention of minorities. In the Netherlands, more specifically, the incorporation of ethnic minorities has become the topic of debates around the notions of citizenship and nationhood (Scholten and Holzhacker 2009). These disputes resonate with the old Chicago School assimilationist language in which a simple chain of events was held to be responsible for the accommodation of minorities into the mainstream. As with the old assimilationist debate, integration in the Netherlands is now seen by some policymakers as the end point to the cycle of interethnic contact. Moreover, what is meant by integration is not only participation in socio-economic life, but also adherence to a set of common norms and values (Joppke 2004; Scholten and Holzhacker 2009). Amsterdam is no exception to the attack on multiculturalism (Uitermark, Rossi, and

van Houtum 2005; Vertovec and Wessendorf 2010). Ever since a Dutch man of Moroccan descent murdered Theo van Gogh, a film-maker who fervently criticized Islam, ethnic concentration seems to have become taboo. Civil servants fear subsidizing

Ethnic and Racial Studies, 2015 Vol. 38, No. 1, 158-175, https://dx.doi.org/10.1080/01419870.2013.826811

cultural and religious activities. Instead the local government promotes diversity under the rubric of enhancing ‘contact’ between majority and minority populations.What risks to be forgotten is that ethno-national organizations might still be contributing to the (political) integration of new and old immigrants – albeit ‘by a detour’ (Berger, Galonska, and Koopmans 2004). The Chicago School was criticized precisely because it ignored differential routes that minorities take into, for example, the labour market, housing market, the educational system and so on. (Portes and Zhou 1993; Alba and Nee 2005). Not necessarily because minorities do not want to mingle with the majority population, but sometimes because they face discrimination or because they otherwise lack the language skills to participate in social life. In line with the American literature, this differential route is labelled as segmented assimilation by European researchers and acculturation is not seen as the only viable integration route (Vermeulen 2010). To be sure, ethno-national organizations were initially promoted by the local government to facilitate integration (Penninx and Slijper 1999) rather than the result of purposeful self-segregation by minorities. In what follows below, I compare a segmented1 form of participation in voluntary

organizations with a diverse setting that is now favoured by some policymakers and sections of the Dutch public. I question the assumption that ethnic concentration in voluntary or non-profit organizations should a priori be regarded as problematic and take the view that we should put the ‘contact hypothesis’ in voluntary organizations empirically to the test. This article thus investigates the effect of the ethnic composition of voluntary organizations on generalized trust and poses the following questions: to what extent does the level of generalized trust differ across Turkish and mixed organizations, and to what extent does it differ across their participants? In addition, do members of mixed organizations self-select into those organizations on the basis of their prior trust levels or are they socialized into trusting individuals due to participation? The findings suggest that participants of Turkish organizations have overall less

generalized trust than participants of mixed organizations. However, I demonstrate that participants of mixed organizations in Amsterdam might self-select into hightrusting organizations, controlling for their length of participation.2 In addition, the data suggest that generalized trust is consistently better explained by having a higher vocational qualification or a university degree as compared to having no qualifications or having finished primary school. A final finding is that generalized trust seems to be lower among participants who have experienced divorce or been widowed. Before discussing the results, I will elaborate on the voluntary organizations and

generalized trust nexus and which factors generate generalized trust. I focus on the contact mechanism at the organizational level, although many studies have recently focused on ethnic diversity at the neighbourhood level. The research strategy of this study overcomes shortcomings of previous research by oversampling individuals in organizations (see Stolle 1998; Maloney, van Deth, and Roßteutscher 2008). Moreover, voluntary organizations are a small enough setting to investigate the effects of faceto-face contact between ethnically diverse groups on generalized trust. Using multilevel modelling and ordinary least squares regression, I then test if generalized trust is affected by participation in Turkish and mixed voluntary organizations in Amsterdam. Apart from contact between ethnically diverse groups, social success factors can also