Measuring value-added in higher education: Possibilities and limitations in the use of administrative data

This paper develops a general methodology for measuring the value added of institutions of higher education using commonly available administrative data. Our approach recognizes the data limitations and selection problems inherent in higher education, and highlights the challenges these issues pose for education policy. Combining information from different administrative sources in the state of Texas, we follow the universe of Texas college applicants from the time of application (pre-enrollment) through public college and into the labor market. In specifications that do not control for selection, we find large, significant differences across colleges in terms of persistence, graduation, and earnings; however, these differences decrease substantially when we control for selection. In light of the growing interest in using value-added measures in higher education for both funding and incentivizing purposes, our methodology offers unique evidence and lessons for policy makers.