Ethnic Diversity and Social Trust: Mapping a Diverse Field through Meta-analysis

• In Event: Post-Conflict Dynamics

Abstract

Background: Does ethnic diversity erode social trust? Answering this question is crucial for modern mass societies that are predicated on trust between strangers, and which are currently becoming increasingly ethnically diverse due to immigration. Yet, despite an abundance of empirical analyses, little agreement over the effects of ethnic diversity on social trust has been reached. The purpose of this paper is to summarize and systematize the findings on this relationship through a meta-analysis of existing empirical studies.
Method: We conducted a meta-analysis based on a data set consisting of more than 1500 estimates of the relationship between ethnic diversity and various forms of social trust from previous studies. For each estimate we have coded a host of substantive (e.g. geographical setting), methodological (e.g. research design, measurement of trust and ethnic context, inclusion of potentially confounding and moderating variables), and disciplinary (e.g. field in which a study is published) features of the study. This enables us to map variations in the results across each of these features and thereby systematically address a number of hypotheses proposed in the literature including the effects of diversity provided contact or contact-prone conditions.
Results: Tentative results show a modest negative effect of ethnic diversity on trust across all studies, but also important variations in this relationship based on various features of the studies. The relationship appears to be stronger when examined in more local contexts as predicted by theories emphasizing interethnic exposure as the underlying mechanism. The negative relationship exists for all types of trust, but is strongest for trust in neighbors, thus highlighting the experiential foundation of this form of trust. Finally, the strength of the relationship between ethnic diversity and trust is sensitive to model specifications in terms of choice of control variables. Interestingly, widely discussed confounders, such as socio-economic deprivation, turn out as far less important than purported mediators of the relationship.
Conclusions: Summarizing previous studies through a meta-analysis, we find a modest negative effect of ethnic diversity on trust, but with substantial and theoretically meaningful variation across studies. We conclude the review by discussing potential avenues for future research that would increase theoretical and methodological sophistication in the field.
Tentative results show a modest negative average effect of ethnic diversity on trust across all studies, but also important variations in this relationship based on various features of the studies. The relationship appears to be stronger when examined in more local contexts as predicted by theories emphasizing interethnic exposure as the underlying mechanism. The negative relationship exists for all types of trust, but is strongest for trust in neighbors, thus highlighting the experiential foundation of this form of trust. Finally, the strength of the relationship between ethnic diversity and trust is sensitive to model specifications in terms of choice of control variables including purported mediators of the relationship.

Authors

• Peter Thisted Dinesen, University of Copenhagen

• Kim Mannemar Soenderskov, Aarhus University