How Robust are Country Rankings in Educational Mobility?

Abstract

As robustness of social science is getting more attention, analytical choices have turned out to be more important than previously thought. In this paper, we investigate the robustness of intergenerational educational mobility estimates using multiverse analysis, a technique for incorporating many analyses into one framework while varying analytical choices such as variable coding, mobility measures, and exclusion criteria. Using European Social Survey (ESS) data from 16 European countries, results show substantial variation, both in point estimates and rank orderings among countries. Even so, the central tendency of country ranks accords well with previous literature. While any single ranking comes with considerable uncertainty, the field has nevertheless produced a relatively robust set of empirical findings.