Wiswall, M., Stiefel, L., Schwartz, A., & Boccardo, J. (2014). Does attending a STEM high school improve student performance? Evidence from New York City. Economics of Education Review, 40, 93-105.
- The study’s objective was to examine the impact of New York City (NYC) STEM high schools on math and science test-taking behavior and scores.
- The study used regression analyses to compare outcomes for students who attended a STEM-focused high school with those who attended a regular public high school. The study used administrative data from the NYC Department of Education.
- The study found that young women attending STEM high schools in NYC were more likely to take various New York State Regents Examinations in math and science and to score higher on biology tests compared with young women who attended non-STEM public high schools.
- The quality of causal evidence presented in this study is low because the treatment and comparison groups were not equivalent on key measures at baseline. This means we are not confident that the estimated effects are attributable to STEM high schools; other factors are likely to have contributed.
New York City (NYC) STEM High Schools
Features of the Intervention
STEM high schools in NYC teach specific science and math programs to every student. Admission to these schools is based on a number of criteria, including student preference, school-specific admission criteria, and space availability. In addition to offering this specialized curriculum, STEM high schools employ different teachers than do the regular public high schools—on average, NYC STEM teachers have more experience teaching and are more likely to have earned a master’s degree than their counterparts at regular public high schools. Teachers at these schools also earn higher salaries than those in the traditional public schools. Because the study cannot disentangle the effects of the STEM programs themselves and the different teacher characteristics, this review considers STEM teachers to be a component of the treatment condition as implemented in this study.
Features of the Study
The study used data from high schools in the NYC public school system, which included 30 STEM schools. There were 181 non-STEM schools in 2007 and 227 in 2008 included as comparisons. The authors used regression analyses to compare STEM with non-STEM schools on both the probability of taking New York State Regents Examinations in various subjects (Math A, Math B, Chemistry, Physics, and Biology) as well as the scores received on those tests. Students in New York must pass a series of five tests in order to earn a Regents high school diploma, including at least one math and one science exam, or can earn an Advanced diploma by passing additional exams (a non-Regents diploma can be earned by passing any one exam). The analyses controlled for middle school scores in math and reading, race and ethnicity, free lunch eligibility, English language use at home, limited English proficiency, and age. The authors ran regression models by gender to estimate gender gaps both within and across school types.
- The study found that female students in NYC STEM schools were significantly more likely to take Math B, Chemistry, and Physics Regents exams and to earn higher scores on the Biology exam compared with female students in non-STEM schools.
Considerations for Interpreting the Findings
Students in NYC chose whether to attend a STEM or non-STEM school. The authors were unable to control for unobservable differences between the two types of students, including potential selection factors such as parental support and student motivation, that might influence a student’s decision regarding which type of school to attend. Thus, estimated effects might reflect differences in outcomes that are due to these differences in student characteristics as well as the impact of the STEM school.
The study measured the effectiveness of STEM high schools in NYC. The results might not be externally valid in other cities, particularly if other STEM schools do not share the same characteristics as the NYC STEM schools, including offering higher teacher salaries and hiring more educated and experienced teachers.
Causal Evidence Rating
The quality of causal evidence presented in this report is low because the treatment and comparison groups were not equivalent on key measures at baseline. This means we are not confident that the estimated effects are attributable to NYC STEM schools; other factors are likely to have contributed.