Improving the Measurement of the STEM-Trained Population using Generalized Method of Moments
Yu-Chieh Hsu, University of Chicago and NORC
Janna E. Johnson, University of Minnesota
Javaeria A. Qureshi, University of Illinois at Chicago
Policymakers have recently emphasized the need to improve science, technology, engineering, and math (STEM) education in the United States to ensure the future competitiveness of the country’s workforce in an increasingly technology-dependent world economy. To design policies targeting the STEM workforce, accurate measures of its size and diversity are essential. Current measures of the size of this relatively small population, derived from national-level surveys, vary widely. To improve measures of the STEM workforce, we apply a Generalized Method of Moments (GMM) estimator that combines datasets from multiple sources. We show that this method yields much more accurate measures of the number of individuals with college majors in specific STEM fields in 1993 and 2003 than what is found using survey data alone. The GMM estimates produce very different measures of gender and racial gaps in these majors than what is shown in the raw survey data.
Presented in Poster Session 5: Economy, Labor Force, Education and Inequality