Can societies be identified displaying the proposed form of social cohesion with all
the positive qualities it is said to include? It is important to explore this since we need
to know whether some proposed version of social cohesion refers to an actual real-life
phenomenon or merely to a hypothetical state of affairs. If the latter is the case,
policies devised to enhance social cohesion are likely to fail, for instance because they
benefit only some constituent components and have unintended negative side effects
for other components.
This study will therefore explore whether one or several phenomena can be
observed empirically that combine components associated with social cohesion and
that could thus be labeled as real-life manifestations of social cohesion. Additionally,
it will assess to what extent these real-life manifestations conform to perspectives that
I labeled “universalist” and “particularist” (see explanation in the third section). I start
by explaining the distinct approach followed by this study and by reviewing the main
theoretical approaches to social cohesion that identify it as a macro-level
phenomenon. The purpose of this exercise is to identify a collection of empirically
observable components of social cohesion suggested by these approaches. I then
discuss the universalist and particularist perspectives and explain which forms of
social cohesion would support either of these perspectives. Subsequently, I explain
the data sources, the indicators selected to tap the social cohesion components and the
methods of analysis. Finally I present the main findings.
Social cohesion: assessing macro-level approaches
Interestingly, even most of the empirical studies of social cohesion have not critically
engaged with the multidimensional conception of social cohesion advanced by many
theoretical approaches. These studies have by and large proceeded from one
theoretical approach and explored how the concepts offered by such an approach can
be made measurable. Examples of such studies are Letki (2008) and Rajulton et al
(2007), who developed measurements of social cohesion as a local phenomenon, and
Dickes et al (2009), who relied on Bernard’s (1999) theoretical approach to construct
indicators of social cohesion as a societal feature. By not challenging the theoretical
models, these studies merely reproduce the multidimensional conceptions of social
cohesion advanced by these models. Rajulton et al (2007), for instance, have
developed a composite index of social cohesion composed of three domains for
census metropolitan areas (CMAs) in Canada. They remark that CMAs can
compensate a low score on one domain with a high score on the two others and thus
still achieve a relatively high ranking on social cohesion. They miss the point however
that the social cohesion rankings thus obtained represent qualitatively different forms
of social cohesion. CMAs with the same score on the social cohesion index can have
very different social cohesion profiles, while CMAs with different scores may show
quite similar profiles. These likely outcomes put the usefulness of composite
multidimensional indexes into question.
To my knowledge so far only Green et al (2003; 2006) have explored to what
extent the social cohesion components suggested by various theoretical approaches
co-vary. Their analysis which focused on national level characteristics produced a
syndrome of social cohesion consisting of aggregate levels of social and institutional
trust, civic compliance and (the absence of) violent crime. However, apart from
highlighting social capital theory, they did not specify the theoretical approaches that