This paper aims to test different approaches of human capital stock approximation. It faces one of the main questions in explaining link between human capital and economic growth. It tries to step forward in answering what is the best proxy of human capital. It starts from Barro & Lee and Cohen & Soto datasets which are expanded by Mincerian approach to human capital measurement and educational structure of population as a human capital proxy. The original dataset covering 73 countries within 1960-1990 is being re-tested and results from panel data regression analyses are compared with expanded dataset.