How to select students for advanced or elite academic programs has long been controversial. Critics of “holistic” admissions policies argue they often turn to mush—or inject bias into the process. At the other extreme, a few programs use nothing more than a single assessment to determine placement. Most notably, gaining entrance to New York City’s prestigious specialized public high schools requires the prospective student excel on a single opt-in standardized test, called the SHSAT exam.
A new working paper examining high school admissions policies in Mexico City sheds light on the limitations of one-shot tests like the SHSAT and suggests that more information about an applicant will help schools and students to be better matched.
In the new study, researchers from Tulane University and VIA Education, a Mexican non-profit, use achievement and demographic data from middle schoolers in Mexico City to explore assignment to the city’s public high school. The paper aims to answer two critical questions: Do students admitted to elite schools perform better? And does it make sense to admit students based on a single test?
In Mexico City, students applying to high school put their top choices on a list, and then they are assigned to schools based on both the preferences they listed and their scores on an admissions test. Since the elite schools are oversubscribed (i.e., more students apply than they can accept), the researchers are able to calculate the effect of getting admitted to one of these elite schools for students near the threshold.
Using a regression discontinuity framework, the researchers compare students who gained admission to elite schools to students who just missed the test score cutoff. Overall, there was no effect on high school graduation rates for students near the threshold of elite school admission. Yet they also found that the effect of attending an elite high school—for those students at the margin—depended on whether the students had good grades in middle school. Students with strong middle school grades saw a 7-percentage-point increase in graduation rates when admitted to elite schools. In contrast, those with worse middle school grades experienced an 8-percentage-point decline in graduation rates when placed in elite institutions.
This suggests that using grades in conjunction with the score on the test provides better predictions of future student performance. An admissions system that considers only a single metric may be ignoring valuable information that could be used to create better matches between students and schools. In addition, the students who perform well on tests may not accurately represent the overall population of high achievers, at least when achievement is defined more broadly.
The study went on to conduct simulations of Mexico City’s elite school composition if middle school grades were weighted alongside exam scores. The results showed that, under this broader approach, the selected students would be less affluent and more likely to be girls compared to the current test-only approach. This finding suggests that incorporating multiple indicators of student performance not only leads to better predictions of student success, but that it could also lead to more equitable and diverse educational opportunities in Mexico City’s high schools—or under any admissions system that is based on narrow criteria.
SOURCE: Marco Pariguana and Maria Elena Ortega-Hesles, “School Choice, Mismatch, and Graduation,” University of Western Ontario Working Paper (November 2022).
