Association for Public Policy Analysis & Management

Panel: The Role of Administrative Data in Informing Higher Education Policy

(Education)

In a chapter prepared for inclusion in a forthcoming volume of the Handbook of the Economics of Education, David Figlio and coauthors review the myriad ways in which computing advances, coupled with increased access to administrative data sets, have enabled education researchers and practitioners to conduct rigorous, high-impact, policy-relevant research. However, the majority of this research, and the associated gains in knowledge, have occurred in the context of primary and secondary (K-12) education. Much remains to be learned about the post-secondary educational production function and the efficacy of post-secondary interventions. This panel begins to fill this gap by exclusively focusing on the use of administrative data to conduct research relevant to post-secondary education policy and practice. Consistent with the theme of the 2016 APPAM Fall conference, the papers in this session show how administrative data, quantitative analysis, and computing power can be combined to produce research and knowledge that can be leveraged to make government, public policy, and non-profit institutions operate more effectively and efficiently. The first paper analyzes proprietary administrative data on SAT “score sends” by students to colleges and universities. Previous research has used “score sends” to proxy for applications, yet preliminary results indicate that “score sends” poorly predict actual applications. Moreover, the degree of measurement error systematically varies by student and school type, which raises questions about the validity and interpretation of results that rely on “score sends” as a proxy for applications. This paper exemplifies how administrative data are not a panacea, and must be carefully evaluated and cross-validated. The second paper uses administrative data from the Texas Public University System to study how the realized cost of attending graduate school changed in response to a change in financial aid policy that eliminated graduate students’ ability to take subsidized Stafford loans. Despite making up only 16% of students, graduate students hold about 40% of outstanding student loans. This paper illustrates how policy impacts can be evaluated using post-secondary administrative data from a statewide system. The final two papers use administrative data from specific postsecondary institutions, and exploit natural experiments within those institutions, to further our understanding of the post-secondary education production function and identify malleable policy levers that improve the achievement, attainment, and labor market outcomes of under-represented groups. These studies exploit the (quasi) random assignment of mentors and peer advisors to students to isolate the causal effect of such support on student persistence, performance, and field of study. Together, the papers in this panel span an array of methodological and policy issues relevant to the study of postsecondary educational policy, as well as the potential benefits and pitfalls of using post-secondary administrative data. The session organizer, chair, discussant, and policy reactor all have significant experience analyzing and interpreting educational administrative data, which will ensure both specific feedback to the paper presenters as well as a fruitful, general discussion on how researchers, practitioners, and policy makers can address the many importance challenges facing higher education access and policy today, and in the future, using administrative data.