Economics of Higher Education

Paper Session

Friday, Jan. 6, 2017 10:15 AM – 12:15 PM

Hyatt Regency Chicago, Grand Suite 5
Hosted By: American Economic Association
  • Chair: Susan Dynarski, University of Michigan

Influencing Along Observables: Estimating College Students' Sorting Into Peer Groups and Its Educational Impact

Jakina Debnam
Cornell University


Using a novel set of behavioral social network data which captures online "friendship" and messaging of undergraduate students at an elite northeastern university, I estimate the role of sociodemographic characteristics – gender, race, first-generation status and citizenship – in peer network formation over time. Using exponential random graph modeling, I find evidence that the role played by sociodemographic characteristics – gender, race, first-generation status and citizenship – in network formation when students' relationships with one another are exclusively virtual differs from the role played by these characteristics when relationships may include a face-to-face component (e.g. when students are on campus).

The Path to College Education: Are Verbal Skills More Important Than Math Skills?

Esteban Matias Aucejo
London School of Economics and Political Science
Jonathan James
California Polytechnic State University


The aim of this paper is to study the differential roles of math and verbal skills for educational outcomes. By estimating a multi-period factor model of skills, using a rich panel database that follows all students in England from elementary school to university, we find that verbal skills play a greater role in explaining university enrollment than math skills. In addition, we use our framework to study the timing of skill development during compulsory schooling. Results show that 40% of skills measured at the end of compulsory education are developed between the ages of 7 and 16, which indicates some scope for overcoming initial skill disadvantages. Finally, we study the gender gaps in college enrollment and STEM field enrollment, showing that verbal skills and comparative advantage in skills are key determinants of these gaps.

Does Classroom Diversity Improve Academic Outcomes?

Yan Lau
Reed College


This paper estimates the causal effect of racial diversity in the classroom on academic outcomes. I exploit a quasi-experimental setting where first-year students in a year-long mandatory humanities writing course at a U.S. liberal-arts college are assigned to discussion conference groups with varying levels of diversity in terms of classmate racial composition. This within-classroom diversity is effectively random conditional on students’ scheduling availability, given the institutional features determining conference assignment, and the fact that students do not know (ex-ante) and cannot manipulate the racial composition of peers in their enrolled conference. I find that a higher degree of racial diversity in the conference causes a statistically significant increase in the humanities course grade and the grade point average (GPA) at graduation. However, diversity has no statistically significant effect on GPA at the end of the first year. These results contribute to the debate over affirmative action in higher education, and offer modest justification for race-based admissions policies.

Is Post-Secondary Education a Safe Port and for Whom? Evidence From Canadian Data

Diana Alessandrini
Auburn University


Previous studies document that adverse labor market conditions, proxied by the unemployment rate, stimulate post-secondary enrollment. This paper shows for the first time that aggregate unemployment not only affects total enrollment but also changes the composition of the student body and students’ educational path, with important consequences for inter-generational mobility. Contrary to previous research, I use Canadian data from the Survey of Labor and Income Dynamics (1993-2011). The advantage is twofold. First, Canada has a more extensive post-secondary education sector compared to other countries. Second, the panel structure of the dataset allows to study the impact of labor market conditions beyond initial enrollment. Results show that unemployment stimulates university enrollment especially among individuals with highly educated parents. This has consequences for educational inequality. Individuals with highly educated parents are also more likely to choose university over college when unemployment rises. Thus, labor market conditions affect the type of education and skills that students acquire. Further, aggregate unemployment has a nontrivial impact on the decision to drop out of school and the decision of workers to return to school.

Scientific Education and Innovation: From Technical Diplomas to University STEM Degrees

Nicola Bianchi
Northwestern University
Michela Giorcelli
Stanford University


This paper uses a change in enrollment requirements in Italian STEM majors to study the effects of university STEM education on the probability of becoming an inventor. Administrative data on education, occupations, and innovation activities of students who received a STEM degree thanks to the change in enrollment policy suggest that the propensity to innovate decreased among students with high pre- collegiate achievement, but increased among lower-achieving students. We show how these findings relate to heterogeneous sorting into more and less innovative occupations. In addition to affecting occupational choices, a university STEM education changed the type of innovation produced.
JEL Classifications
  • I2 - Education and Research Institutions