Paper presented at: Warwick AMES, 30th Annual Congress of the European Economic Association,XXIV Meeting of the Economics of Education Association, 2nd Applied Econometrics IAAE, AUEB Conferences in Economics and Econometrics, 6th Annual Meeting of the French Economic Association, 18th IZA European Summer School in Labor Economics, 2014 and 2015 Conferences of the Eastern Economic Association, LEER Workshop on Education Economics, 2015 Royal Economic Society Conference, 2014 International Atlantic Economic Association, Warwick CWIP, University of Essex, Bogazici University, Norwegian School of Economics, North Carolina at Chapel Hill, University of New South Wales, University of Queensland and University of Cyprus.
We study the effect of disclosing relative performance information (feedback) on students' performance in high-school, on subsequent university enrollment, and on expected subsequent earnings. We exploit a large-scale natural experiment in which students in some cohorts receive information about their relative performance within their schools and across the nation. Using unique primary data, we find an asymmetric response to feedback: high-achieving students improve their final-year performance by 0.15 of a standard deviation, whereas the final-year performance of low-achieving students drops by 0.3 of a standard deviation. The results are more pronounced for females, indicating greater sensitivity to feedback. We also document the long-term effects of feedback: high-achieving students reduce their repetition rate for the national exams; they enroll into university departments that are more prestigious by 0.15 of a standard deviation and their expected annual earnings increase by 0.17 of a standard deviation. By contrast, the results for low-achieving students are negative. We provide suggestive evidence that feedback encourages more students from low-income neighborhoods to enroll in university and to study in higher-quality programs, which may, in the long run, reduce income inequality.
Keywords: feedback, relative performance, university admission, rank, gender differences, income inequality
2) “Persistency in Teachers' Grading Biases and Effect on Longer Term Outcomes: University Admission Exams and Choice of Field of Study" with Victor Lavy
Presented at: 2017 NBER summer institute, University of Queensland, University of Warwick, 2018 RES conference in Sussex (scheduled), 2018 LEER conference in Leuven (scheduled), 28th Australia New Zealand Econometric Study Group Meeting, 2018 AEA conference in Philadelphia, 2017 Asian and Australasian Society of Labour Economics, 2017 Natural Experiments in History: Development, Health, and Labour at Deakin U, 2017 Public Choice Conference at Deakin U, 2017 Applied Econometrics Workshop at Victoria U of Wellington, research seminar at UTS, 2017 Labour Econometrics Workshop at Auckland, 2017 conference for the 80th Anniversary at Belgrade, 2018 ACE in Sydney,
Recent research focus on what shapes gender differences in academic achievements and in university field of study. In this paper we focus on how teachers’ gender role attitudes and stereotypes influence the gender gap by affecting the environment at school. We explore the extent to which teachers’ gender bias in high school influences students’ academic performance in high-stake exams that determine admission to universities and on students’ choice of university field of study. We use data from large number of high schools in Greece where the performance in these high-stake exams determine university admission. We measure teachers’ bias as the difference between a student’s school exam score in 11th and in 12th grade (scored ‘non-blindly’ by the students’ teachers) and her national exam score (taken at the end of 11th and 12th grade and scored blindly). We then define a teachers’ bias measure at the class level by the difference between boy’s and girl’s average gap between the school score and the national score. Positive values indicate that a teacher is biased in favor of boys in a particular subject. We link teachers over time and are therefore able to get a persistent teacher bias measure based on multiple classes, and the effect is estimated for later students’ performance. The panel data on teachers relieves concerns that our measure of gender bias may just pick up random (small sample) variation in the unobserved "quality" or "non-cognitive" skills of the boys vs. girls in a particular single class or any other class specific dynamics. Our results may be summarized with three broad conclusions. First, the same teachers who are biased for one class are biased in the same way for other classes in the same year and in classes in earlier or later academic years. The very high correlations of within teachers’ biases in different classes reveal high persistency in teachers’ stereotypical behavior. Second, teachers’ biases in core and elective subjects (classics, social science, science, exact science) have positive effect on boys’ and negative effects on girls’ performance on end of high school university admission exams. Female teachers are more pro-girls on average but the effect of female and male teachers’ biases on national exams are not statistically different. Third, teachers’ biases in specific courses lower the likelihood that students enroll in a related field of study at the university. This average effect masks large heterogeneity by gender, being larger and statistically significant for girls and not different from zero for boys.
Keywords: teacher bias, gender discrimination, stereotypes
3) "Which degrees do students prefer during recessions? ” Accepted, Empirical Economics
with Sofoklis Goulas
Featuring in: The Conversation, and Epoch Times (Singapore), Scrollin (India) , EUROPP (LSE), The Kathimerini (in Greek) and The Ethnos ( in Greek)
This paper examines the changes in higher education demand that took place in Greece over the period 2005-2011, with a focus on the role played by the increase in unemployment rate on the demand for different fields of university study. We use administrative data on the number of applications submitted to each undergraduate program in Greece, combined with a degree specific job insecurity index and time series on youth (18-25 year old) unemployment. Results indicate that the steep increase in unemployment rate that started in 2009 was associated with an increase in the number of college applicants. The effect is heterogeneous across fields, with an increase in the demand for degrees in Psychology as well as for Naval, Police and Military Academies, and a decrease in the demand for degrees in Business and Management. We find that job insecurity turns applicants away from degrees that are associated with poor employment prospects. We also find that the business cycle affects degrees' admission thresholds.
Keywords: demand for education, college major, unemployment, job insecurity, admission thresholds
4) "Swine Flu, Class Attendance, and School Performance. Should we force students to go to class? (Under Review) with Sofoklis Goulas
presented at: 2018 RES Conference in Sussex, 2017 RES PhD Conference, 2016 SAEe Simposio de la Asociacion Espanola de Economia in Bilbao, 31st Annual Congress of the European Economic Association in Geneva, 2016 Conference in Economic Theory and Econometrics, AUEB in Tinos, 4th International PhD Meeting of Thessaloniki in Economics in Thessaloniki, 65th Annual Meeting of the French Economic Association in France, 2016 Eastern Economic Association, Annual Conference in Washington DC, University of North Carolina. Scheduled: Monash University, Health Economics Institute
In this paper, we investigate the effects of the choice to skip class on scholastic outcomes. We exploit exogenous variation from a natural experiment in Greece that relaxed the time budget constraint of high school students to identify the effect of absences on scholastic outcomes across the ability distribution. In the school year 2009-2010 high school students were allowed to skip 30% more hours of class in comparison to previous or following years with no penalty. This treatment was provided to protect students from the Swine Flu. Using data on swine flu cases we provide evidence that the treatment affected absences but not school performance directly. We use an instrumental variable approach to identify the intention-to-treat effect of the policy as well as well-defined local average treatment effect of absences on grades. We find that the relaxed class attendance policy caused an increase in absences of roughly 10 hours. In addition, we find a positive effect of absences upon grades.
with Sofoklis Goulas
Abstract:Would you prefer a tighter or a prolonged exam schedule? Would you prefer to take Math before Reading or the other way around? We exploit variation in end-of-course exam schedules across years and grades to identify distinct effects of the number of days between exams, the number of days since the first exam, and the exam order on subsequent performance. We find substantially different scheduling effects between STEM and non-STEM subjects. First, we find a positive relationship between exam performance in STEM subjects and exam order, controlling for other influences of scheduling, suggesting that the later in the schedule an exam is taken the higher the average performance. We call this phenomenon, exam warm-up. Second, we find a negative relationship between the number of days from the very first exam and subsequent exam performance in STEM subjects, suggesting the existence of a fatigue effect. For STEM subjects, the fatigue effect is estimated to be less than half the size of the warm-up effect. For non-STEM subjects, an additional day between exams is significantly associated with lower performance in subsequent exams. Students of lower prior performance have lower fatigue effects and higher warm-up effects in STEM subjects compared to students of higher prior performance. Also, we find that exam productivity in STEM increases faster for boys than it does for girls as they take additional exams due to a higher warm-up effect. Our findings suggest that low-cost changes in the exam schedule may have salient to student performance gaps.
Work in Progress
Lecturer in Economics
“Does the girl next door affect your academic outcomes and career choices? ” with Yi Zhang and Sofoklis Goulas
Status: Draft in preparation. Paper presented at: 1) CAGE seminar, University of Warwick, 2) Brown Bag seminar, University of Queensland, 3) PhD meeting in Economics, University of Macedonia, July 2016, 4) CSDA Workshop, University of Auckland and 5) Monash, Department of Economics, Research seminar and 6) Conference in Economics and Econometrics, University of Athens (AUEB). Scheduled: European Population Conference 2018 in Brussels, XXVII Meeting of the Economics of Education Association in Spain
Peer effects are potentially important for optimally organising schools and neighborhoods. In this paper, we examine how the gender of classmates and neighbors affects a variety of high school outcomes and the choice of major at the university level. Given that students are assigned to schools based on proximity from residential address, we define as neighbours all same-cohort peers who attend any other school within a mile from one's school. To control for potentially confounding unobserved characteristics of schools and students that might be correlated with peer gender composition, we exploit within-school and -neighborhood idiosyncratic variation in the proportion of females across consecutive cohorts in the twelfth grade. Using data for the universe of students in public schools in Greece between 2004 and 2009, we find that a higher share of females in school or neighborhood improves both genders' subsequent academic performance, increases the university admission rate for both genders, makes both genders more likely to enrol in university compared to education institutes and affects the choice of university study. In particular, we find that only females are more likely to enrol in STEM university departments and earn higher wages when they have more female peers in school or neighborhood. Our effects are of higher magnitude for girls compared to boys and we also find that the effects coming from neighbors are almost as large as the ones coming from schoolmates. We also find that: 1) the effect is non-linear, namely the effects are larger for schools and neighborhood cohorts with a large majority of female peers and 2) the effects are larger in low-income schools and neighborhoods.
“Feedback, Learning and Performance: Evidence from a Randomised Controlled Trial" with Isabella Dobrescu, Marco Faravelli and Alberto Motta
Status: Draft in preparation, Slides available upon request
We report results from a large scale RCT to investigate the impact of feedback on students’ relative performance in a research intensive, selective university. Students are provided with information on their relative ranking in a semester long computerized assignment. The RCT encompasses five treatments varying the timing and nature of the feedback. We find evidence that feedback is more effective if it is provided only when relative position is affected. In particular, we find that the provision of this type of feedback led to an increase of around 4% in students’ grades. We also find that the effect was significant for the whole distribution and both types of exams (e.g. midterms and final). Feedback biased toward positive or negative variations in ranking is less effective.
Keywords: feedback, randomised controlled trial, ranking, college students, learning
"The impact of scholarships and bursaries on academic success in University and the Labor Market”
Status: Results ready, New Draft is coming soon
Abstract: This paper examines the benefits to the University students of entry scholarships and bursaries and identifies the effects on the University graduation performance, completion time and the labor market. Students enrol in Universities based on the admission’s score comprised of high school national exams and national rank. I use unique data from the Hellenic Public Scholarship Department combined with student level performance data directly collected from University Departments. Each year the Top1 % of the University entering students in each Department gets an automatic merit award of around €1200. Although studying at the University is free, this award acts mainly as a motivational device. Using a regression discontinuity analysis, I find that scoring top in the entrance exams improves students’ academic performance throughout the system, increases the probability to get a bursary next year, increases the probability to get a Master’s scholarship and has a significant positive effect on the probability to move abroad in order to find a job. Using data from a Tax Authority, I find that one out of three students who get access to this flow of financial resources leave the country, enhancing the brain-drain effect. Effects are stronger for students coming from low-income neighborhoods.
Keywords: regression discontinuity, merit aid, college enrollment
“May I be excused? Identification of returns to absences and class peer effects” with Sofoklis Goulas
Abstract: In this paper, we investigate (1) returns to absences and (2) peer effects. We exploit exogenous variation from a natural experiment that changed the school absences allowance for the better students in order to identify the effect of school attendance on educational outcomes. The natural experiment took place in Greece in 2007 and provided higher performing students with 50 more hours of excused absences from school. We start off by using a Regression Discontinuity approach in order to measure the change in total absences and exam score due to the reform around the cutoff. Next, we employ a combination of differences-in-differences and instrumental variables techniques in order to identify returns to absences and peer effects. Our estimates show significant positive peer effects in Greek Language but negative peer effects in Mathematics. Furthermore, our estimates yield significant negative returns to absences of a magnitude of 0.01 standard deviations per hour of absence in Greek Language, Mathematics and the overall GPA.
Keywords: human capital, returns to education, attendance, peer effects, natural experiment
“Does the Gender Composition Matter in Justice? Evidence from the Supreme Judicial Court”
“Class size and Performance” with Kala Krishna and Sergey Lychagin
"The Effect of Peer Effects and the choice of STEM degrees" with Yi Zhang and Sofoklis Goulas