Higher Education

May 1, 2009
Charles Clotfelter, Organizer

Yona Rubinstein, Brown University, and Sheetal Sekhri, University of Virginia
Do Public Colleges in Developing Countries Provide Better Education than Private Ones? Evidence from General Education Sector in India

In many developing countries, public colleges outperform private colleges. Graduates of public colleges do far better than their private counterparts as manifested in their college exit examination outcomes. This often has been attributed to the cutting edge education provided in public colleges. However, public colleges are highly subsidized, suggesting that the private-public education outcome gap might reflect the pre-determined quality of the students who sort into public colleges, rather than the causal impact of the public tertiary education on student's outcomes. Rubinstein and Sekhri evaluate the impact of public colleges using a newly assembled unique dataset that links admission data with the educational outcomes on a set of common exit exams in India. Admission to general education public colleges is strictly based on the results of the Senior Secondary school examinations. The researchers exploit this feature in a regression discontinuity design, and find that the public colleges have no added value at the margin of selection. Controlling for entry scores, they find no differences between the exit exam outcomes of students graduating from public and private colleges.


Jason Lindo and Nicholas Sanders, UC, Davis, and Philip Oreopoulos, University of British Columbia and NBER
Ability, Gender, and Performance Standards: Evidence from Academic Probation

Lindo, Sanders, and Oreopoulos use a regression discontinuity design to examine students’ responses to being placed on academic probation. Consistent with a model of introducing performance standards, they find that being placed on probation at the end of the first year discourages some students from returning to school while improving the performance of those who return. The researchers find heterogeneous responses across ability, gender, and native language, and extend the model to explain why low ability students are not more discouraged than high ability students. They also analyze the effects on GPAs, suspension, and graduation rates


Philippe Aghion, Harvard University and NBER, Mathias Dewatripont, Université Libre de Bruxelles, Caroline Hoxby, Stanford University and NBER, Andreu Mas-Colell, Universitat Pompeu Fabra, and Andre Sapir, Université Libre de Bruxelles
The Governance and Performance of Research Universities: Evidence from Europe and the U.S.

Aghion and his co-authors investigate how university governance affects research output, measured by patenting and international university research rankings. For both European and U.S. universities, the researchers generate several measures of autonomy, governance, and competition for research funding. They show that university autonomy and competition are positively correlated with university output, both among European countries and among U.S. public universities. Next, the researchers identity a (political) source of exogenous shocks to the funding of U.S. universities. They demonstrate that when a state's universities receive a positive funding shock, they produce more patents if they are more autonomous and face more competition from private research universities. Finally, the authors show that during periods when merit-based competitions for federal research funding have been most prominent, universities produce more patents when they receive an exogenous funding shock. This suggests that routine participation in such competitions hones research skill.


Scott Carrell and Marianne Page, UC, Davis and NBER, and James West, United States Air Force Academy
Sex and Science: How Professor Gender Perpetuates the Gender Gap

Why aren't there more women in science? Female college students are currently 37 percent less likely than males to obtain a bachelor's degree in science, technology, engineering, and math (STEM), and comprise only 25 percent of the STEM workforce. Carrell and his co-authors exploit a unique dataset of college students who have been randomly assigned to professors over a wide variety of mandatory standardized courses. The researchers focus on the role of professor gender.

Their results suggest that while professor gender has little impact on male students, it has a powerful effect on female students' performance in math and science classes, their likelihood of taking future math and science courses, and their likelihood of graduating with a STEM degree. The estimates are largest for female students with very strong math skills, who are arguably the students most suited to careers in science. Indeed, the gender gap in course grades and STEM majors is eradicated when high performing female students' introductory math and science classes are taught by female professors. In contrast, the gender of humanities professors has only minimal impact on student outcomes. These results may be indicative of important environmental influences at work.


Shulamit Kahn, Boston University, and Megan Macgarvie, Boston University and NBER
How Important is U.S. Location for Research in Science?

Kahn and MacGarvie ask whether scientists located outside the United States are at a disadvantage when it comes to research productivity. Using a new and unique dataset of foreign-born U.S.-educated scientists that allows them to exploit exogenous variation in post-Ph.D. location induced by visa status, they are able to compare students who were required by law to leave the United States upon the completion of their studies with similar students who were allowed to remain in the United States. The authors assess whether the students who left the United States have more or fewer publications, citations, and collaborators, as compared to a control student with the same advisor. Research output is defined as the number of publications, first-authored publications, publications in high-impact journals, and the publications’ impact on science as measured by the number of forward citations. Using visa status to instrument for location, and allowing richer and poorer countries to have different impacts, the authors find that the negative relationship between non-U.S. location and research output is present and large for poorer countries but completely eliminated when the researcher is located in a richer country, with two one exceptions. Foreign location negatively affects last authorship and collaboration with Americans, even for those located in the richest countries. Further, allowing for heterogeneity in the treatment effect of foreign location on research output on these same countries, Kahn and MacGarvie find that the negative effect on publications of being abroad is largest for those with the lowest estimated propensity of being abroad, those who – given their observable characteristics -- would be expected to remain in the United States.


Katja Maria Kaufmann, Stanford University
Understanding the income Gradient in College Attendance in Mexico: The Role of Heterogeneity in Expected Returns to College

When examining reasons for low school attendance, researchers face an important identification problem: on the one hand, people might expect low returns to schooling and thus decide not to attend; on the other hand, they might face high attendance costs that prevent them from attending, despite high expected returns. To address this identification problem, Kaufmann uses data on people’s subjective quantitative expectations of future returns to schooling, which can be shown to affect their schooling decisions. With data on Mexican high school graduates, she can analyze the causes and implications of the steep income gradient in college enrollment in Mexico. Her data on people’s expected returns and on their schooling decisions allow her to directly estimate and compare cost distributions of poor and rich individuals. She finds that poor individuals require significantly higher expected returns to be induced to attend college, implying that they face higher costs than individuals with wealthy parents. She then tests predictions of a simple model of college attendance choice in the presence of credit constraints, using parental income and wealth as a proxy for the household’s (unobserved) interest rate. She finds that poor individuals with high expected returns are particularly responsive to changes in direct costs such as tuition, which is consistent with credit constraints playing an important role. Evaluating potential welfare implications by applying the Local Instrumental Variables approach of Heckman and Vytlacil (2005) to her model, she finds that a sizeable fraction of poor individuals would change their decision and attend in response to a reduction in the interest rate. Individuals at the margin have higher expected returns than the individuals already attending college, which suggests that policies such as governmental student loan programs could lead to large welfare gains.