ABSTRACT
The alarming difference in the percentage of confirmatory findings between the disciplines of economics, political science and sociology demands for an interdisciplinary approach. It suggests that the overall inconclusiveness of the debate is the function of a missing consensus about appropriate methodological set ups, ranging from the operationalization of diversity and affected outcomes, to suitable control variables in estimation procedures. A starting point for any future interdisciplinary debate should acknowledge that the overall mixed evidence, with a tendency toward confirmatory findings, does not question whether the relationship actually holds, but suggests that it depends on moderating conditions. In order to formulate hypotheses under which conditions ethnic diversity should negatively affect public goods production and social cohesion, we need to have a deeper understanding and empirical tests of the mechanisms that we believe to drive the association. Social cohesion has a behavioural component in the forms of civic infrastructure, and in membership in associations and other engagements in public social life.
