![Sample our Politics & International Relations journals, sign in here to start your FREE access for 14 days]({"type":"image" , "src" : "/sda/5947/TOC-Subject-Sample-2021-Politics-International-Relations.jpg"})
Abstract
Scholars of political disaffection have recently argued that ethnically bridging and bonding non-profit associations may have diverging effects on members’ likelihood of political engagement. However, both theoretical and empirical assessments of such effects remain inconclusive by predicting different scenarios. In this article, I contribute to this literature by relying on a novel design, which does not focus on individuals (as in previous research), but instead samples ethnically homogeneous and heterogeneous organisations and their participants. This allows partitioning variance in political engagement at these two levels of analysis, and admits a more direct test of the link between ethnic composition of an organisation and political engagement. This sampling strategy also allows controlling for the effects of previously ignored organisational characteristics, which may enhance political engagement. A multilevel analysis suggests existing studies tend to overestimate the importance of an ethnically heterogeneous organisational context. Findings suggest that both types of voluntary organisations attract politically engaged individuals (self-selection). Moreover, other organisational characteristics—such as providing political information or receiving (local) government grants—have a stronger effect on members’ political engagement than an organisation’s ethnic composition.
Acknowledgements
The author gratefully acknowledges suggestions from Benny Geys, Bertjan Doosje, Floris Vermeulen, Tom van der Meer, Tim Reeskens, Matthijs Vastenburg, Eva-Maria Trüdinger and two anonymous reviewers of this journal.
Disclosure statement
No potential conflict of interest was reported by the author.
Notes
[1] Participants usually filled out a questionnaire in Turkish, Dutch or English depending on their fluency in either of these languages. If they requested to be contacted later, I would send them a link to a web-based questionnaire. Occasionally, when the board found that my visit would disrupt an event, I circulated questionnaires via them. I instructed the board member to give the questionnaire to a diverse set of participants or to send an email invitation to all participants.
[2] However, later I had to introduce a snowball method to complement the stratified random sample. This was because some organisations were disbanded or due to non-response the sample would not contain enough mixed organisations. Moreover, the sampling frame did not include recently established organisations. The snowball method is based on information from the Internet and from informants in the city council and other organisations. It may be argued that introducing this selection method would result in sample bias. However, I have aimed at a stratified sample that would include different types of organisations. Moreover, when working with a small sample, here 20 Turkish and 20 mixed organisations, a purely random sampling frame may result in a non-representative sample. Therefore, the sample here is representative of a key situation, namely bridging and bonding ethnic ties, and includes the same number of sports associations, women’s groups, etc. in each type.
[3] Since the participants did not know about my presence at the events, there is no reason to expect them opting out. It is certainly the case that people may have all sorts of reasons and motivations to participate in one event, but not in another, but there is no reason to believe that their political engagement levels may systematically have affected their participation in the event that I visited.
[4] This paper does not intend to make any claims about the civic and political engagement of the Turkish population—including its linguistic and religious minorities. Such a fine-grained analysis should be based on a probability sample including information on various ethnic and religious groups who emigrated from Turkey to the Netherlands or Amsterdam more specifically.
[5] Voting is sometimes regarded as an undemanding activity. The question thus remains does the likelihood of voting (or vote intention) rule out other political activities. Empirical analyses show otherwise: ‘a citizen who would do one of these activities would probably do others as well’ (Norris Citation2002, 7). In a recent study, Marien, Hooghe, and Quintelier (Citation2010, 198) find that self-reported voting behaviour ‘is a very frequent activity’ across 25 European countries, whereas non-institutionalised forms of political participation are much less frequent.
[6] This scale performs very well as RMSEA value is zero (Kline Citation2011). However, the Chi-square statistic for this test is significant (Stevens Citation2002). But an exploratory factor analysis model shows very high factor loadings as well as a high Cronbach alpha value (α = 0.933).
[7] The CFA model performs moderately as the Chi-square statistics is significant. However, the RMSEA value is zero. Cronbach alpha is reasonable as well with this small number of items (α = 0.594), while the factor loadings are high.
[8] Entering the original diversity score for each organisation in the models below does not change the results.
[9] Since ethnic background correlates highly with generational differences, I cannot add both variables to the models. Therefore, future research should focus on any additional effects generational differences may have on political engagement. Nevertheless, as Bloemraad (Citation2013) suggests once we have controlled directly for all other factors, which are seen to be the drivers of political engagement, generational differences may not be substantive.
[10] In a series of additional interaction effects (results available upon request) I gauged whether length of membership in conjunction with either political discussion, confidence in government or political interest had significant effects. Similarly, I assessed whether mobilisation activities of organisations interacted with the length of participation, and whether participants of Turkish organisations, which offered more mobilisation activities or received grants showed higher political engagement levels. None of these interactions were statistically significant. In addition, I collected data on multiple membership and voluntary work. Adding these variables to the models does not statistically change the results, which goes to suggest that additional (active) memberships do not incrementally increase the likelihood of engagement (see Van der Meer and Van Ingen Citation2009). All these robustness analyses should assure us that there are no additional socialisation effects in general or specifically arising from participation in bonding associations due to mobilisation. Nor do socialisation in other organisations than surveyed here relate to political engagement.