The accumulation of the human capital stock plays a key role to explain the macroeconomic performance across regions. However, despite the strong theoretical support for this claim, empirical evidence has been not very convincing, probably because of the low quality of the data. This paper provides a robustness analysis of alternative measures of human capital available at the level of EU NUTS1 and NUTS2 regions. In addition to the univariate measures, composite indicators based on different construction principles are proposed. The analysis shows a significant impact of construction techniques on the quality of indicators. While composite indicators and labour income measures point to the same direction of impact, their correlation is not overwhelmingly high. Moreover, popular indicators should be applied with caution. Although schooling and human resources in science and technology explain some part of the regional human capital stock, they cannot explain the bulk of the experience.