Inequality with Ordinal Data. Cross-Disciplinary Review of Methodologies and Application to Life Satisfaction in Europe
Ordinal data are ubiquitously available and frequently employed in empirical research. A review of methodological approaches from various scientific fields shows that in different fields of the social sciences rather distinct methods for measuring inequality are applied without appreciating works in other disciplines. Synthesizing these works leads to the proposal of a research design, which combines dominance criteria and inequality measures from different families. This design is applied to the most recent data on life satisfaction in Europe, a field of research continuing to gain political importance but typically building on questionable methodologies. Hence, empirical results presented are valuable in several respects. One important finding is that specific measures are not consistent as re-rankings frequently occur. It seems this is not primarily due to the underlying data being ordinal, but more because dominance can hardly be found. Indeed, the underlying 0-10 Likert-scale appears to cause a dilemma as the Allison-Foster principle of dominance turns out to be restrictive when applied to relatively dense scales.
SCHODER, Jörg (2014) Inequality with Ordinal Data. Cross-Disciplinary Review of Methodologies and Application to Life Satisfaction in Europe. University of Salzburg, Department of Geography and Geology, Geographies of Uneven Development – Working Paper, 12/2014, No. 5