Meetequivalentie in internationaal vergelijkend onderzoek

Abstract

It is indisputable that comparative research contributes to sociological knowledge by providing insight in the differences that exist across national and cultural contexts. However, valid cross-national comparisons require theoretical constructs to be measured equivalently across countries. This is especially the case for abstract concepts, such as values, attitudes, or opinions. As such, measurement equivalence cannot be readily assumed; it is a hypothesis that needs to be tested empirically. This article reviews the social science literature on the cross-national comparability of measurements. We start with some conceptual clarifications regarding the central notion in this field, namely ‘measurement equivalence’. Possible sources of inequivalence as well as preventive measures are discussed. A subsequent section deals with statistical models to test empirically whether the conditions for measurement equivalence are fulfilled. Most attention is paid to the most popular technique for testing equivalence, namely multiple group confirmatory group analysis (MGCFA). By means of illustration, we test whether the ESS-scale measuring support for the welfare state is comparable across respondents from the Netherlands, Flanders and Wallonia. Finally, we suggest what may be done when equivalence is not supported by the data.