Abstract: The stock of human resources available to a society is integral to its economic growth and development. Unfortunately, as a composite of levels of embodied human capital and accumulated experience, its measurement and comparison across societies is hampered by its inherently latent nature. Usually, for the purpose of analysis, some form of Cantril scale is arbitrarily attached to the ordered categorical variable proxies of educational status and age group, but this is problematic since results can be ambiguous when using scale dependent summary statistics for comparison purposes. Here, new scale independent techniques for making inferences about the respective levels of, and differences between, human resource stocks across groups are proposed that are not subject to such concerns. Their effectiveness is exemplified in an application comparing Canadian Provincial Human Resource Stocks in the 21st century.
Keywords: Human Resource Measurement, Ordered Categorical Data, Stochastic Dominance
JEL Classification: O15; I00; J01, C14