Measuring cultural diversity within countries with ellipses, contour maps, and distances

Abstract

Cross-country cultural analyses compare different aggregations of national culture where a country is represented by a single data point. Using the Integrated Values Survey (1981 - 2022), we explore various methods to quantify cultural diversity within each country. First, we show how principal components (PCs) are extracted such that the PCs retain individual-level information consistent with country-aggregate representation, which better accounts for variance in the data. On the resulting culture maps, we then use standard deviational ellipses (SDE) and contour maps (CM) to measure national population diversity. We also use vectors of cultural traits to measure pairwise Euclidean distance between individuals within the same country. Cultural diversity rankings were then compared across these methods, to find that they are poorly correlated (Spearman correlation -0.33-0.27) with ethnic, linguistic, and religious fractionalization (FR), indicating additional information accounted for in our diversity rankings. The diversity measures are easily be incorporated into predictive spatio-temporal cross-cultural analyses that use similar, if not the same culture vector representation, and for analyzing errors and uncertainty.