1.Background

One common feature of education systems worldwide is the significant investment of time and resources by students seeking admission to selective high schools. Numerous studies have explored the impact of attending selective schools on academic performance, yet the findings have been notably varied. While some studies report positive results, most show no significant effect.

There are several potential explanations for the mixed findings. Firstly, attending a more selective school may offer students better educational resources, the opportunity to study alongside high-achieving peers, and the chance to learn from top-quality teachers, which may have positive effects. Conversely, high-achieving peers can cause students to underestimate their abilities, reducing self-efficacy and leading to poor academic performance. Secondly, inconsistent findings may be attributed to variations in the approaches used to control for selection bias. Some studies have utilized randomized admissions lotteries, which may not be feasible in most educational contexts. As an alternative, other studies have used the discontinuity around the entrance exam cutoff scores. Lastly, differences in education and economic development levels in the researched areas can lead to varying research conclusions even within the same country.

2.Research Questions

Due to the inconclusive nature of previous findings, this study seeks to provide additional evidence on the impacts of selective high schools. We first investigated the effects of attending selective high schools on students’ performance on NCEE. We then divide the selective high schools into two tiers based on their admission scores: Tier-1 high schools (more than 98% of students meet the cutoff for the top 30% four-year universities) and Tier-2 high schools (more than 90% of students meet the cutoff for the top 30% four-year universities). Furthermore, we investigated the effect on college admission, as indicated by the university a student ultimately attends. This outcome is crucial as it reflects not only a student’s academic performance but also a range of non-academic skills, including application strategies, information-gathering abilities, etc. Finally, our research examined the impact on students’ self-efficacy as the potential mechanism. This outcome is important as it strongly correlates with individuals’ academic performance and long-term achievement.

3. Significance

This study makes several significant contributions to the existing literature. One advantage is that we were able to investigate the effects on a broader range of outcomes, such as total scores, test scores for each subject in the National College Entrance Exam (NCEE), and perceived self-efficacy, which offers a more comprehensive understanding of the effects of selective high schools. Furthermore, our study has the advantage of observing the university that students ultimately attend.

Second, our data was collected from the capital city, where the level of education and economic development is among the best in China, and school choice is actively discouraged by the local education department to foster educational equity. The impacts in this context may differ substantially from those observed in other contexts, such as rural counties. Additionally, China introduced a new scheme for NCEE in 2014, which may impact the dynamics of school choice and the effects of school quality. Our data includes students who took NCEE in 2021, enabling us to investigate the impact of selective high schools after implementing the new scheme.

Lastly, our data enabled us to test the impacts of Tier-1 and Tier-2 selective high schools separately, providing a more detailed understanding of the effects of attending these schools.

4.Data and Measures Our data were collected through an online survey administered to high school graduates. The survey incorporated comprehensive information, including students’ demographic characteristics, family background, NCEE scores, college admissions, etc. The survey encompassed responses from 11,495 high school graduates who participated in NCEE and met the four-year undergraduate college admissions criteria. This study only focuses on students whose middle and high schools were located in the same urban district. Our analytical sample consists of 4,366 students, including 867 students from 14 Tier-1 high schools, 945 students from 23 Tier-2 schools, and 2,554 students from 80 regular schools. We focus on three main outcomes: academic performance, college admission, and self-efficacy. To assess academic performance, we use four measures, including the NCEE total score and scores for the three main compulsory subjects. We also assess the probability of students being admitted to first-class universities as a measure of college admission. Finally, we evaluate students’ self-efficacy, which is assessed through ten survey items. 5.Methods To determine the causal effects, we used a regression discontinuity design (RDD) based on the fact that students were admitted to selective high schools if their HSEE scores were above a certain threshold. Our analysis utilized a fuzzy RDD approach, given that the probability of being admitted to selective high schools did not change from 0 to 1 at the cutoff point in our data. The fuzzy RD estimator was constructed by fitting the two-stage local polynomial parameter estimation function. We use an indicator of whether a student’s HSEE score crosses the score threshold for attending a Tier-1/Tier-2 high school as an instrument for the treatment, a dummy variable indicating whether a student attends a Tier-1 or Tier-2 selective high school. We employed the local linear and quadratic functions and utilized MSE-optimal bandwidths. In two-stage regression, the same polynomial degree and bandwidth were selected. For robustness checks, we used CER-optimal bandwidths and other bandwidths suggested by Hahn et al. (2001), Imbens and Lemieux (2008), Ludwig and Miller (2007). We also employed non-parametric kernel-based local linear regression with triangular and Epanechnikov kernels. We also checked the validity of RDD. Firstly, our analyses indicated that passing the Tier-1/Tier-2 school cutoff strongly predicts attendance at a Tier-1/Tier-2 school. Secondly, the standardized HSEE score was continuous at the cutoff. Lastly, the students’ characteristics, such as gender and parents’ education years, were continuous at the cutoff.

6.Findings Our results showed that students with higher initial HSEE scores tend to achieve higher NCEE scores and have a greater chance of attending top universities. Nevertheless, Tier-1 schools reduced the NCEE total and math scores of students near the cutoff by about 44 and 17 points, respectively. Few estimates of the Tier-2 school effect were significantly different from zero. These results suggested that participating in selective high schools did not improve academic performance for students near the eligibility cutoff. Students admitted to Tier-1 selective high schools performed even worse academically than their Tier-2 counterparts. Moreover, attending selective high schools did not positively affect students’ college admission outcomes. Selective high schools might have a heterogeneous impact on students’ performance with different characteristics due to differences in their adaptability. Heterogeneity was only observed in the Tier-1 school effect on NCEE math scores of students with different parents’ education years and the Tier-2 school effect on NCEE math scores and first-class university admission of students with different genders. Tier-1 schools significantly reduced NCEE math scores for students whose parents had 16 or more years of education while having no significant impact on those whose parents’ education years were less than 16 years. Tier-2 schools had a significant negative effect on NCEE math scores for female students but an insignificant impact on males. Regarding the probability of attending first-class universities, Tier-2 schools had a significant negative effect on male students but an insignificant impact on females. One possible explanation for the zero to negative effects on both academic performance and college admission was that the coursework difficulty level in selective high schools may be too challenging for some students and decrease their self-efficacy, ultimately leading to unfavorable academic outcomes. However, our results indicated no impact of attending selective high schools on self-efficacy.

Anderson, Kathryn, Xue Gong, Kai Hong, and Xi Zhang. 2016. “Do Selective High Schools Improve Student Achievement? Effects of Exam Schools in China.” China Economic Review 40: 121-34. Barrow, Lisa, Lauren Sartain, and Marisa de la Torre. 2020. “Increasing Access to Selective High Schools through Place-Based Affirmative Action: Unintended Consequences.” American Economic Journal: Applied Economics 12, no. 4: 135-63. Berkowitz, Daniel, and Mark Hoekstra. 2011. “Does High School Quality Matter? Evidence from Admissions Data.” Economics of Education Review 30, no. 2: 280-88. Dee, Thomas and Xiaohuan Lan. 2015. “The Achievement and Course-Taking Effects of Magnet Schools: Regression-Discontinuity Evidence from Urban China.” Economics of Education Review 47: 128-42. Ding, Weili and Steven F. Lehrer. 2007. “Do Peers Affect Student Achievement in China’s Secondary Schools?” The Review of Economics and Statistics 89, no. 2: 300-12. Dustan, Andrew, Alain de Janvry, and Elisabeth Sadoulet. 2017. “Flourish or Fail?: The Risky Reward of Elite High School Admission in Mexico City.” Journal of Human Resources 52, no. 3: 756-99. Elsner, Benjamin and Ingo Eduard Isphording. 2017. “A Big Fish in a Small Pond: Ability Rank and Human Capital Investment.” Journal of Labor Economics 35, no. 3: 787-828. Estrada, Ricardo and Jérémie Gignoux. 2017. “Benefits to Elite Schools and the Expected Returns to Education: Evidence from Mexico City.” European Economic Review 95: 168-94. Hoekstra, Mark, Pierre Mouganie, and Yaojing Wang. 2018. “Peer Quality and the Academic Benefits to Attending Better Schools.” Journal of Labor Economics 36, no. 4: 841-84. Jackson, C. Kirabo. 2010. “Do Students Benefit from Attending Better Schools? Evidence from Rule-Based Student Assignments in Trinidad and Tobago.” The Economic Journal 120(December): 1399-429. Park, Albert, Xinzheng Shi, Chang-tai Hsieh, and Xuehui An. 2015. “Magnet High Schools and Academic Performance in China: A Regression Discontinuity Design.” Journal of Comparative Economics 43, no. 4: 825-43. Pop-Eleches, Cristian and Miguel Urquiola. 2013. “Going to a Better School: Effects and Behavioral Responses.” American Economic Review 103, no. 4: 1289-324. Shi, Ying. 2020. “Who Benefits from Selective Education? Evidence from Elite Boarding School Admissions.” Economics of Education Review 74: 101907. Wu, Jia, Xiangdong Wei, Hongliang Zhang, and Xiang Zhou. 2019. “Elite Schools, Magnet Classes, and Academic Performances: Regression-Discontinuity Evidence from China.” China Economic Review 55: 143-67. Zhang, Hongliang. 2016. “Identification of Treatment Effects under Imperfect Matching with an Application to Chinese Elite Schools.” Journal of Public Economics 142: 56-82.