An empirical assessment of the cross-national measurement validity of graded paired comparisons

The popular use of graded paired comparisons in empirical studies assessing consumers’ preferences, and the potential effect of cross-national differences in (extreme) response styles on the quality of graded paired comparison data, supply ample reasons for an empirical verification of the cross-national validity of such scales. Using data from a cross-national margarine brand study including fourteen different nations (N = 4,560), we found sufficient statistical evidence for cross-national bias due to existing cross-national differences in extreme responses. However, the low values reported for effect size measures (intra-class correlation coefficient, R2 value) indicated that the impact of the cross-national bias is marginal. The findings from our study provided empirical support for the hypothesis that graded paired comparison data can be meaningfully compared across nations