negative loadings of these indicators), and consensus on support for democracy and
gender equality. The core components of this factor are social trust, equality and
disparities on traditional/secular values with loadings of .7 and higher or -.7 and
lower. Similarly, the second factor captures civic participation, ethnic tolerance, lack
of institutional trust, national pride and pluralism on gender equality (which all have
loadings of more than .50 or less than -.50), has civic participation as core component
(i.e. with a loading of more than .70) and accounts for 20 per cent of the variance. The
remaining two factors each comprise just one indicator (political discussions and level
of belonging). Given their insignificance we chose to disregard them in further
analyses.
Table 4 about here
I am thus left with two factors. Given the variety of indicators they comprise it is not
easy to find appropriate labels, but in view of their core components I provisionally
call them „solidarity’ and „participation’. I saved the country scores on these factors.
High scores on the solidarity factor thus represent high levels of social trust, equality,
diversity on religious/secular values, etc; low scores represent their opposite. A
similar logic applies to the participation factor.
Interestingly, the four shared values indicators are related in opposite ways to
the solidarity factor: the higher the score on solidarity, the greater the disparities on
religious/secular and survival/self-expression values and the smaller the disparities on
support for gender equality and democracy. It highlights all the more the need to
distinguish between values in examining how shared values relate to social cohesion.
More intriguingly, it seems to sustain the idea that societies can handle substantial
degrees of cultural pluralism on a plethora of issues provided there is consensus on a
number of key norms and values regulating inter-group conflicts (Parsons 1970). This
somewhat contradicts the opinion put forward by Mann (1970: 423) that social
cohesion does not depend on value consensus but on the “pragmatic acceptance by
subordinate classes of their limited roles in society”. More research into the precise
role of shared values is needed, however, to state this conclusion with more certainly.
Most importantly, however, the factors extracted are not in agreement with
any of the definitions of the aforementioned macro-level approaches. In other words,
these approaches all advance incoherent, multidimensional conceptions of social
cohesion. Remarkably, even the model proposed by Green et al is not in line with any
of the factors even though it is based on empirical analysis (institutional trust is not
correlated with social trust and social order, as in Green’s model, but with tolerance
and civic participation; the inconsistency with the findings of Green et al is probably
explained by the fewer number of countries in their study).
The fact that the aforementioned conceptions of social cohesion do not
represent coherent syndromes does not rule out the possibility that the social cohesion
profiles of specific countries are in line with these conceptions. This would be the
case if these countries exhibit all the components of social cohesion proposed by these
conceptions. As Figure 1 shows, there are indeed four countries - Sweden, Denmark,
Netherlands and to a somewhat lesser extent Finland - that have high values on both
the solidarity and participation factor and that are thus very likely to combine high
levels of social trust, equality, order, sexual tolerance and consensus on basic values
with high levels of participation, ethnic tolerance and national pride. The social