out to have significantly more tolerant views on immigrants the more diverse their
classrooms were, controlling for all relevant individual and classroom conditions. This is
broadly in line with American research showing that desegregation helps to combat racial
prejudice among whites.
However, classroom diversity was not found to be related to ethnic tolerance in
England. Clearly, despite the micro-environment of the classroom meeting all the
conditions that contact theory assumes crucial for diversity to enhance tolerance, this
positive effect need not occur in all national contexts. From this point of view the degree
of support for the contact perspective is perhaps disappointing.
Diversity did not show a consistent link with intended participation across the
three countries either. While it had a significant positive effect in Sweden, it had no
impact in Germany and England. This brings us to the second overall conclusion: the
effect of diversity differs markedly not only across countries but also across civic
outcomes. In other words, any relationship found between diversity and some civic
outcome in - say - America need not apply in other western states, nor can it be assumed
that diversity is related in the same way across different civic outcomes within one
country. Thus, any generic effect of diversity postulated by some theory is not so strong
that it overrules the effect of country-specific factors or can be generalized to other
desirable social outcomes.
An important task for future research is to assess whether one or several common
factors can account for the country variations in the effect of diversity or whether we
have to come to terms with the idea that irreducible, nationally unique configurations of
conditions fundamentally shape the impact of diversity at the school or classroom level.
This results of this study suggest that the latter may well be the case. Obviously, the
policy implication is that policy makers have to be very cautious in borrowing and
implementing a straight copy of a successful education policy on - say - community
cohesion from a different state, however close to one’s own country this state is in
political tradition and culture.
I have to end with one important limitation. I have essentially performed a cross-
sectional analysis using a single point in time database. This raises the issue of direction
of causality and selection effects. In theory it is possible that more tolerant ethnic
18