Care is required when making assertions about the relationship between diversity and social cohesion

Democratic Audit UK recently published a piece by Gal Ariely on the relationship between social cohesion and ethnic and religious diversity. Ariely argued that the the answer to whether the latter is eroded by the former depends on how diversity is conceptualized and measured. Here, Tom van der Meer responds to Ariely, arguing that great care should be taken in investigating this question.

Credit: Chris Ford, CC BY NC 2.0

As cross-national immigration increased sharply after the 1960s, advanced western societies have become ever more diverse. Policy makers and social scientists responded immediately after a 2006 presentation by Robert Putnam at Uppsala University, Sweden. He argued that ethnic diversity and ongoing immigration erode social cohesion. Putnam’s constrict claim poses that faced with an ethnic diverse living environment, people tend “to hunker down – that is to pull in like a turtle”, or in common language, to retreat from social life.

Putnam’s conclusion were cause for concern and therefore received wide attention in the media and amongst policy makers. The very premise of multiculturalism – various ethnic groups living together – was at stake. Scholars rushed forward to test whether the constrict proposition held in a broad range of countries (from the US to Sierra Leone), at a broad range of environments (from neighbourhoods to countries), using a broad set of diversity measures (from outgroup size to linguistic heterogeneity), and on an even broader set of social cohesion indicators (from national pride to volunteering and social trust). The recent study by Gal Ariely is one of those studies.

This led to a cacophony of empirical findings: there are about as many studies that found support for the constrict claim as there are studies that found none. A plurality of studies even came to mixed conclusions. Despite heated scholarly debates, the fundamental question remained unanswered: Is ethnic diversity harmful to social cohesion?

To deal with this cacophony I – together with Jochem Tolsma (Radboud University Nijmegen) recently performed a review study of the literature, published last year in the Annual Review of Sociology. Restricting ourselves to core indicators of social cohesion (social trust, associational involvement, informal help, social ties), we ended up with 90 unique studies covering mostly western countries. Most of these studies covered the US (18), the UK (11), the Netherlands (10), and Canada (7), or are cross-nationally comparative studies (21).

We looked for consistent patterns in the wide variety of conclusions. First, let us first note two patterns that we did not find. 1: The variety of findings is not caused by some studies being qualitatively better than others in terms of methodological set-up. 2: unlike the findings in the study by Ariely, we found no evidence that the type of diversity matters. In some studies it does, in others it does not; even in those studies where it did matter, we found no consistent patterns. Most intriguingly, however, a study by Finseraas and Jakobsson from 2012 argues that diversity is only harmful when various cleavages align, i.e., when the distinction between ethnic groups and religious groups or income groups shows considerable overlap.

What then explains the cacophony of empirical findings? Three patterns stand out.

First, ethnic heterogeneity does not undermine ties (trust, contacts) between ethnic groups. If anything, increased interethnic contact opportunities stimulate interethnic contact, while interethnic contacts stimulate trust.

Second, ethnic diversity does seem to drive down one aspect of social cohesion, namely those ties that are themselves bounded by the neighbourhood, such as trust in and contact with neighbours. However, these harmful effects do not spill over to more general forms of social cohesion such as generalized trust, volunteering and informal help. This distinction between intra-neighbourhood social cohesion and other ties solves most of the debate in the UK and the Netherlands.

Third, the only exception to this pattern is made up by the United States, where we found more evidence that heterogeneity also erodes other aspects of social cohesion. There seems to be a case of American exceptionalism that is commonly suggested in the social capital literature.

All in all, I agree with Ariely that caution is required by scholars and policy makers when dealing with the relationship between ethnic diversity and social cohesion. Diversity only drives down trust in and contact with our neighbours, but not our general civic outlook or inclination to volunteer. Given the speed with which alarming claims were originally embraced by journalists and policy makers, these conclusions are by themselves rather sobering.

—

Note: This article gives the views of the authors, and not the position of Democratic Audit UK, nor of the London School of Economics. Please read our comments policy before commenting.

—

Tom van der Meer is associate professor at the Department of Political Science of the University of Amsterdam, the Netherlands. His research interests cover political trust, social cohesion, and electoral behaviour.