Most parents have strong views regarding which schools to send their children to. Indeed, in many nations, there are preferred or elite public secondary schools for which there is high demand and fierce competition. This is a bit of a puzzle, however, because evidence from various countries shows that attending sought-after public secondary schools does not improve secondary school examination performance. Using meta-analytic methods, Figure 1 shows that the precision-weighted average effect on test scores across all publicly available studies using quasi-random assignment to either a preferred or elite (non-charter) public school (either through lottery or selective enrolment exam cut-offs) is statistically indistinguishable from zero.
Figure 1 Effects of preferred and elite public secondary schools on test scores
Notes: The individual standardised estimated effects along with their 95% confidence intervals are shown. RD denotes a regression discontinuity design exploiting admission scores cut-offs. Admission Lottery refers to identification strategies exploiting randomised admission lotteries for oversubscribed schools. The resulting overall weighted average effect shown is 0.017sd with a 95% prediction interval of [-0.026; 0.059].
One potential explanation for this pattern is that parents’ school preferences are based on different outcomes than secondary school exam performance. If so, preferred schools may improve outcomes valued by parents (such as non-cognitive skills, social skills, job referral networks, adult earnings, and wellbeing) that are not well-measured by test-score impact. In a recent paper (Beuermann and Jackson 2019), we test this hypothesis.
Our study is the first to estimate the causal impact of attending a preferred public secondary school on test scores and also on a broad array of medium- and longer-run outcomes for the same population of students.1 We do so by exploiting rich administrative and survey data from Barbados.
The Barbados data and context are well-suited for this study. At the end of primary school, students take the Barbados Secondary School Entrance Examination (BSSEE). At BSSEE registration, students submit a ranked list of preferred secondary schools to the Ministry of Education, Science, Technology and Innovation (METI), and the METI uses a deferred acceptance algorithm to assign students to schools based on their choices and their test scores.
The assignment rule used by the METI creates a test-score cut-off for each school above which student applicants are admitted and below which they are not. Since the scores for those who just missed and those who just made the cut-off are nearly identical, their underlying abilities are nearly identical. Thus, any difference among those who score above the cut-off is because that group was more likely to attend a preferred school while the other wasn’t. This feature allows us to employ a regression discontinuity (RD) design to identify the causal effect of attending a preferred school.
We use administrative data on the BSSEE, and all secondary-school applications and assignments for 25 years (1987 through 2011). To track educational outcomes, we merge these student-level BSSEE data to administrative school-exam records taken at the end of secondary and post-secondary studies between 1993 and 2016. To track a rich set of long-run outcomes, we link the administrative BSSEE records to the 2016 Barbados Survey of Living Conditions and focus on cohorts aged 25 or older at the time of the survey.2
We show that attending a preferred school is associated with higher-achieving peers, more academically homogeneous peers, and smaller cohorts. However, consistent with previous evidence (across several national contexts), we find no improvement in secondary-school exam performance.
Looking at medium- and longer-run outcomes tells a different story from the short-run test-score impact. Indeed, students at preferred schools are more likely to earn a post-secondary credential. Furthermore, attending a preferred school is associated with more years of completed formal education among individuals between the ages of 25 and 40.
Consistent with the educational attainment effects among these older individuals, persons who attended a preferred school were less likely to be in the labour force between the ages of 17 and 24, but more likely to be engaged in post-secondary studies during those ages. However, the longer-run educational gains are observed only for females.
Consistent with the education patterns, attending a preferred school has no effect on earnings among males, but does increase female earnings. An exploration into mechanisms for women reveals that the earnings increase for women is likely mediated by their being employed in higher-status occupations (as opposed to being more productive at the same job), and we provide evidence that the improved social networks at preferred schools may facilitate securing these higher-status jobs.
To further help explain the gender differences, we examine fertility. Attending a preferred school is associated with reduced teen motherhood but no change in total fertility. This teen-motherhood effect may explain why women at preferred schools are more likely to be in post-secondary studies between the ages of 17 and 24, attain more years of education by age 25, and have higher earnings, while there is no such effect for men. It can also explain the sizable long-run benefits for women despite no test score effects. To our knowledge, this is the first evidence of a causal link between school quality and teen motherhood.
Finally, we also found that both women and men experienced significant long-term health benefits, as measured by a healthy body mass index, regular exercise, and regular dental check-ups, if they gained admission to the schools that their parents preferred. These findings suggest that preferred schools may promote productive habits and attitudes that are not measured by test scores but contribute to overall wellbeing. This may represent a significant, previously undocumented return to school quality.
The fact that we find no impact of attending preferred schools on short-run test scores but do find a sizable impact on short-run non-cognitive outcomes (like teen motherhood) that may trigger effects on longer-run outcomes (like educational attainment and labour market success) is important. Our results suggest that test-score impact may not be the best measure of a school’s impact on longer-run outcomes (Heckman et al. 2006, Jackson 2018). Accordingly, policymakers should be cautious regarding using test score impact in accountability systems and incentive-pay schemes.
Moreover, these results underscore the need for evaluations of school-choice programs and private-voucher programmes to move beyond test-score impact alone and to examine broader sets of outcomes that parents may value.
Abdulkadiroglu, A, J D Angrist and P A Pathak (2014), “The elite illusion: Achievement effects at Boston and New York exam schools”, Econometrica 82(1): 137–196.
Abdulkadiroglu, A, J D Angrist, Y Narita, P A Pathak and R A Zarate (2017). “Regression discontinuity in serial dictatorship: Achievement effects at Chicago’s exam schools”, American Economic Review 107(5): 240–45.
Ajayi, K F (2015). “Student performance and the effects of academic versus nonacademic school attributes”, mimeo, Boston University.
Anderson, K, X Gong, K Hong and X Zhang (2016). “Do selective high schools improve student achievement? Effects of exam schools in China”, China Economic Review 40: 121-134.
Barrow, L, L Sartain and M de la Torre (2017). “The role of selective high schools in equalizing educational outcomes: Using place-based affirmative action to estimate heterogeneous effects by neighborhood socioeconomic status”, Federal Reserve Bank of Chicago.
Beuermann, D W, and C K Jackson (2019). “The short and long-run effects of attending the schools that parents prefer”, NBER Working Paper 24920.
Bui, S A, S G Craig and S A Imberman (2014). “Is gifted education a bright idea? Assessing the impact of gifted and talented programs on students”, American Economic Journal: Economic Policy 6(3): 30–62.
Clark, D (2010). “Selective schools and academic achievement”, The B.E. Journal of Economic Analysis & Policy 10(1): 1–40.
Clark, D, and E Del Bono (2016). “The long-run effects of attending an elite school: Evidence from the United Kingdom”, American Economic Journal: Applied Economics 8(1): 150–176.
Cullen, J B, B A Jacob and S Levitt (2006). “The effect of school choice on participants: Evidence from randomized lotteries”, Econometrica 74(5): 1191–1230.
Dee, T, and X Lan (2015). “The achievement and course-taking effects of magnet schools: Regression-discontinuity evidence from urban China”, Economics of Education Review 47: 128–142.
Deming, D J (2011). “Better schools, less crime?”, The Quarterly Journal of Economics 126(4): 2063–2115.
Dustan, A, A de Janvry and E Sadoulet (2017). “Flourish or fail?”, Journal of Human Resources 52(3): 756–799.
Heckman, J J, J Stixrud and S Urzua (2006). “The effects of cognitive and noncognitive abilities on labor market outcomes and social behaviour”, Journal of Labor Economics 24(3): 411–482.
Hoekstra, M, P Mouganie and Y Wang (2018). “Peer quality and the academic benefits to attending better schools”, Journal of Labor Economics 36(4): 841–884.
Jackson, C K (2010). “Do students benefit from attending better schools? Evidence from rule-based student assignments in Trinidad and Tobago”, The Economic Journal 120(549): 1399–1429.
Jackson, C K (2018). “What do test scores miss? The importance of teacher effects on non-test score outcomes”, Journal of Political Economy 126(5): 2072–2107.
Lucas, A M, and I M Mbiti (2014). “Effects of school quality on student achievement: Discontinuity evidence from Kenya”, American Economic Journal: Applied Economics 6(3): 234–263.
Pop-Eleches, C, and M Urquiola (2013). “Going to a better school: Effects and behavioral responses”, American Economic Review 103(4): 1289–1324.
 Clark and Del Bono (2016) examine the longer-run effect of attending one of three elite schools in Scotland but do not examine whether the long-run impact relate to school impact on short-run test scores.
 By age 25, 99% of all individuals had completed their formal schooling. Therefore, this population is suitable to measure educational attainment and labour market outcomes.