Many studies have demonstrated a negative correlation between cultural diversity and trust and cooperation, but they have left the question of the causal nature of this relationship largely open, and the mechanisms behind it unexplored. This large project was intended to fill some of these gaps by conducting a comparative analysis across three countries, across cities, and in the context of schools, using both survey and experimental data. The "Ethnic Diversity and Collective Action Survey" (EDCAS; see Schaeffer et al. 2011 for the technical report) was conducted across local populations in 67 cities and towns in Germany, France, and the Netherlands with a total sample size of 9100 respondents. The survey especially focused on neighbourhood trust and other measures of social cohesion, as well as interethnic networks and civic engagement. We also conducted a survey of parents of children in Berlin primary schools. In both surveys, experiments were integrated – priming experiments in both the German EDCAS and the school survey, and a cooperation game in the school survey. In Berlin, we also conducted a field experiment on cooperative behavior in which we investigated the return rates of letters that were apparently lost on the pavement. Two dissertations were completed on the basis of the project, both graded with highest honours (Schaeffer 2014; Veit 2014). The various experiments make a key contribution by providing consistent and strong evidence that ethnic diversity challenges cooperation. The two priming experiments (Koopmans & Veit 2014; Veit 2015) demonstrate that raising the salience of ethnic diversity leads to more negative judgments of cooperation, both with regard to peoples’ trust in their fellow neighbors to return a lost wallet, and with regard to parents’ judgments of parental cooperation at school. Raising the salience of diversity in age or income, by contrast, has no such effect. The cooperation experiment (see Veit 2014a) reveals less successful coordination of efforts to produce a public good among parents in more diverse schools and classrooms. The lost-letter field experiment (Koopmans & Veit 2014b), finally, demonstrated that return rates of letters are lower in ethnically diverse neighbourhoods, controlling for other neighbourhood characteristics. This experiment also tests the most-often proposed mechanism behind reduced cooperation in diverse contexts: in-group favouritism. We find no evidence for this mechanism because letters are returned more often from homogenous neighbourhoods dominated by native ethnics regardless of whether the addressee is German, Christian, Muslim, or Turkish. By contrast, all letters are returned less often from more diverse neighbourhoods where many Turks and Muslims live. Results of the EDCAS survey confirm the negative relationship between diversity and trust, as well as collective efficacy, and experienced neighbourhood social problems. They also provide further insight into the mechanisms behind this relationship. In one study (Koopmans & Schaeffer 2015a) we focus on perceptions of the quantity and various aspects of the qualitative nature of diversity and test three explanations that have been proposed in the literature for negative diversity effects: out-group biases, asymmetric preferences and coordination problems. We show that all three mechanisms matter. Another study shows that children and interethnic partnerships act as brokers of interethnic contacts that mitigate negative diversity effects (Schaeffer 2013a). Finally, the project made significant contributions to the theoretically adequate measurement of diversity (Schaeffer 2013b; Koopmans & Schaeffer 2015b). In the context of the project a large international conference on “Ethnic diversity and social capital: Mechanisms, conditions, and causality” was organized in 2013. A selection of papers was recently published as an edited volume (Koopmans, Lancee & Schaeffer 2015).