Economics of Education
November 13-14, 2014
Atila Abdulkadiroğlu, Duke University;
Joshua Angrist and Parag Pathak, MIT and NBER; and
Peter Hull, MIT
Lottery estimates suggest oversubscribed charter schools boost student achievement in urban districts. But these estimates needn't capture treatment effects for students who haven't applied to charter schools or for students attending charters where demand is weak. This paper reports estimates of the effect of charter school attendance on middle-schoolers in charter takeovers in New Orleans and Boston. Takeovers are traditional public schools that close and then re-open as charter schools. Students enrolled in the schools designated for closure are eligible for "grandfathering" into the new schools; that is, they are guaranteed seats. Abdulkadiroğlu, Angrist, Hull, and Pathak use this fact to construct instrumental variables estimates of the effects of passive charter attendance: the grandfathering instrument compares students at schools designated for takeover with students who appear similar at baseline and who were attending similar schools not yet closed, while adjusting for possible violations of the exclusion restriction in such comparisons. Estimates for a large sample of takeover schools in the New Orleans Recovery School District show impressive gains. In Boston, where the researchers can compare takeover and lottery estimates, takeover charters generate achievement gains as large or larger than the gains for students assigned seats in lotteries.
Daniel Hungerman, University of Notre Dame and NBER, and
Kevin Rinz, University of Notre Dame
Using a new dataset constructed from nonprofit tax-returns, this paper explores how several recent large-scale choice programs in the U.S. affected the fiscal health of private schools and the accessibility of a private education. Hungerman and Rinz find that school choice programs created a large transfer of public funding to private schools, suggesting that every dollar of funding raised revenue by a dollar or more. Turning to the incidence of school choice and the impact of choice on enrollment, they find that the effects of choice depended crucially on the type of program introduced, with programs based on corporate tax credits creating relatively large enrollment gains and small price increases. The researchers compare results for religious and secular schools, calculate elasticities of demand and supply for private schools, and discuss welfare effects. Back-of-the-envelope calculations suggest that the deadweight loss created by school choice programs is reasonably small: about 5 cents for every dollar of funding.
Joseph Altonji, Yale University and NBER, and
Richard Mansfield, Cornell University
Altonji and Mansfield consider the classic problem of estimating group treatment effects, such as school effects, on an outcome when people sort into groups based on observed and unobserved individual characteristics that affect the outcome. The researchers show that under some circumstances standard choice models imply that group averages of observed individual-level characteristics can serve as controls for all of the across-group variation in the unobservable individual characteristics. This permits estimation of a lower bound for the variance of group treatment effects across groups. The authors use the idea to provide lower bound estimates of the effect of school systems and associated neighborhoods on high school graduation, college attendance, and earnings. Across three data sets, their most conservative estimates indicate that choosing a 90th quantile school and surrounding community instead of a 10th quantile school increases the probability of high school graduation by between 0.047 and 0.085 and increases the college attendance probability by between 0.11 and 0.13. The authors also find large effects on adult earnings. They discuss a number of applications of their methodology, including the measurement of teacher value added.
Wesley Bignell, University of Washington, and Harold Cuffe and Glen Waddell, University of Oregon
While existing research supports that participation in high school athletics is associated with better education and labour-market outcomes, the mechanisms through which these benefits accrue are not well established. Bignell, Cuffe, and Waddell use data from a large public-school district to retrieve an estimate of the causal effect of high school athletic participation on absenteeism. They show that active competition decreases absences, with most of the effect driven by reductions in unexcused absence--truancy among active male athletes declines significantly, with the effects larger in earlier grades and for black and Hispanic boys. Strong game-day effects are also evident, in both boys and girls, as truancy declines on game days are offset with higher rates of absenteeism the following day. Addressing the effects on academic performance, the researchers find significant heterogeneity in the response to active athletic participation by race, gender and family structure, with boys not in dual-parent households exhibiting modest improvements in semesters in which they experience greater athletic participation.
Peter Bergman, Columbia University
In this paper, Bergman studies the long-run impact of a court-ordered desegregation ruling on education outcomes. This ruling mandates that seven school districts, which serve higher-income, predominantly-white families, accept a fixed number of minority elementary school students each year who apply to transfer from a nearby, predominantly minority school district. The fixed number of slots are allocated to families via lottery. The offer to transfer increases the number of students who enroll in college by 7 percentage points. This result is driven by greater attendance to two-year and public colleges, though there are substantial heterogeneous effects. A secondary analysis provides suggestive evidence that peer enrollment matters. Increases in the share of Hispanic students who receive an offer to transfer impacts the likelihood of college attendance among other students who receive the same offer.
Robert Jensen, University of Pennsylvania and NBER, and
Leonardo Bursztyn, University of California, Los Angeles and NBER
When academic effort or investment is observable to peers, students may act to avoid social penalties or gain social favor (i.e., "peer pressure") by conforming to prevailing norms. To test this hypothesis, Jensen and Bursztyn conducted an experiment in Los Angeles high schools in which 11th grade students were offered complimentary access to a commercial, online SAT preparatory course from a well-known test company. Sign up sheets differed randomly across students (within classrooms) only in the extent to which they emphasized that the decision to enroll would be kept private from their classmates. The researchers find that whether choices are believed to be observable to others has dramatic effects on sign up rates. Further, the effects depend greatly on the setting or prevailing peer group norm. In non-honors classes, the sign up rate was 11 percentage points lower when decisions to enroll were to be public rather than private. But sign up in honors classes was unaffected. Since the differential response in the two types of classes could be driven by differences across students in honors and non-honors classes, to further test for the effects of peer pressure the authors examine students taking the same number of honors classes (e.g., the set of students taking exactly two honors classes). For these students, it is essentially random whether the researchers' team arrived and offered them the course during a period in which they were in class with their honors peers or their non-honors peers. When offered the course in their non-honors class, these students were 25 percentage points less likely to sign up if the decision was public. But if they were offered the course in one of their honors classes, they were 25 percentage points more likely to sign up when the decision was public. These results show that students are highly responsive to who their peers are and what the prevailing norm is when they make decisions. These results also allow the authors to isolate peer social concerns from other peer effect mechanisms, since they changed nothing about a student's actual set of peers (or their teachers, classroom or school), only those peers actually present when decisions were made and thus to whom the choice would potentially be reveal.
David Autor, MIT and NBER;
David Figlio, Northwestern University and NBER;
Krzysztof Karbownik, Northwestern University
Jeffrey Roth, University of Florida; and
Melanie Wasserman, MIT
Autor, Figlio, Karbownik, Roth and Wasserman use a new data source matched birth and school records for over one million born in Florida between 1994 and 2002 to investigate the role of family factors in explaining the gender gap in educational outcomes including test scores, truancy, behavioral problems, and high school graduation. The researchers find that boys born to unmarried, low-income, and less-educated mothers have comparatively worse educational outcomes than do their sisters, and importantly, that these differences are not present in birth outcomes. Neighborhood and schooling factors do not appear to drive these effects. The authors conclude that boys are more sensitive to household environments than are girls, and that differences in household advantage explain a substantial portion of the black-white difference in gender gaps in outcomes. Relative to their sisters, black boys fare worse than white boys in substantial part because black children both boys and girls are exposed to worse environments than are white children.
Richard Murphy, University of Texas at Austin, and
Felix Weinhardt, Humboldt University Berlin
In this paper, Murphy and Weinhardt examine the long-run impact of ordinal rank during primary school on productivity using comprehensive English administrative data. Identification is obtained from variation in test score distributions across cohorts and subjects, such that the same score relative to the school mean can have different ranks. Conditional on cardinal measures of achievement, being ranked highly during primary school has large effects on secondary school achievement, with the impact of rank being more important for boys than girls. Using additional survey data the authors find that the development of confidence is the most likely mechanisms for this effect on task-specific productivity.
Esteban Aucejo, London School of Economics, and Teresa Foy Romano, Duke University
While instructional time is viewed as crucial to learning, little is known about the effectiveness of reducing absences relative to increasing the number of school days. In this regard, this paper jointly estimates the effect of absences and length of the school calendar on test score performance. Using administrative data from North Carolina public schools, Aucejo and Foy Romano exploit a state policy that provides variation in the number of days prior to standardized testing and find substantial differences between these effects. Extending the school calendar by ten days increases math and reading test scores by only 1.8% and 0.6% of a standard deviation, respectively; a similar reduction in absences would lead to gains of 5.4% and 2.8% in math and reading. The researchers perform a number of robustness checks including utilizing flu data to instrument for absences, family-year fixed effects, separating excused and unexcused absences, and controlling for a contemporaneous measure of student disengagement. The authors results are robust to these alternative specifications. In addition, their findings indicate considerable heterogeneity across student ability, suggesting that targeting absenteeism among low performing students could aid in narrowing current gaps in performance.
Nathaniel Hilger, Brown University
Intergenerational mobility is a major social objective but its long-term evolution is imperfectly understood. Hilger estimates intergenerational mobility based on children's final schooling in U.S. census data back to 1940. The estimates rely on a novel strategy to account for the large share of children who leave home before final schooling can be observed. Results replicate recent findings of stable mobility since 1980 and lower mobility in the South. However, the results suggest a substantial longer-term increase in mobility, as well as major long-term convergence of mobility in the South to non-South levels. Post-war mobility gains were driven by voluntary high school enrollment, not compulsory schooling laws or college enrollment. Educational gains of poor children relative to rich children plausibly increased aggregate annual earnings growth by 0.125-0.25 percentage points over the 1940-1980 period.
Angela Duckworth, University of Pennsylvania;
John Gabrieli, Amy Finn, and Rebecca Martin, MIT;
Christopher Gabrieli, Harvard Graduate School of Education and Transforming Education;
Matthew Kraft, Brown University; and
Martin West, Harvard University
Duckworth, Finn, Gabrieli, Gabrieli, Kraft, Martin, and West use self-report surveys to gather information on a broad set of non-cognitive skills from 1,368 8th-grade students attending Boston public schools and linked this information to administrative data on their demographics and test scores. At the student level, scales measuring conscientiousness, self-control, grit, and growth mindset are positively correlated with attendance, behavior, and test-score gains between 4th- and 8th-grade. Conscientiousness, self-control, and grit are unrelated to test-score gains at the school level, however, and students attending over-subscribed charter schools with higher average test-score gains score lower on these scales than do students attending district schools. Exploiting charter school admissions lotteries, the researchers replicate previous findings indicating positive impacts of charter school attendance on math achievement but find negative impacts on these non-cognitive skills. The authors provide suggestive evidence that these paradoxical results are driven by reference bias, or the tendency for survey responses to be influenced by social context. The results therefore highlight the importance of improved measurement of non-cognitive skills in order to capitalize on their promise as a tool to inform education practice and policy.
Eric Bettinger, Stanford University and NBER;
Michael Kremer, Harvard University and NBER;
Carlos Medina and Christian Posso, Sr, Banco de la Republica de Colombia;
Maurice Kugler, Harvard University; and
Juan Saavedra, University of Southern California;
A large-scale government program in Colombia used a lottery to distribute scholarships for private secondary school to socially disadvantaged students. Based on administrative data up to seventeen years after the scholarship lottery, Bettinger, Kremer, Kugler, Medina, Posso, and Saavedra document that lottery winners are less likely to repeat grades, more likely to graduate from secondary school on time or ever, and more likely to start and complete tertiary education. Scholarships reduce teen fertility, although there is no significant effect on overall fertility at age 30. Among males, there is some evidence (significant at the 10% level) that winners are less likely to qualify for Colombia's conditional cash transfer program. Point estimates suggest that total formal sector earnings at age 30 are 8 percent greater for lottery winners, a difference that is significant at the 7% level. Impacts on estimated future earnings, including imputed values for those currently in tertiary education suggest a 9.3 percent difference that is significant at the 5% level. Preliminary analyses suggest the expected net present value of increased net tax receipts due to the program exceed the program's fiscal cost, and the program is welfare improving as long as externalities on non-recipients are positive, zero, or negative but less than $1,100 per scholarship recipient.