Association for Public Policy Analysis & Management

Amid concerns over low education standards, watered-down curricula, and consequently skills not rewarded by colleges and employers, 26 states currently require students to pass some form of high school (HS) graduation exams to graduate. These exams can help raise student productivity, better preparing them for college and the workforce, and help signal to potential employers and colleges that a HS graduate has some level of productivity. Graduation exams, however, may come at the expense of increasing inequality in the education system if subgroups of students are disproportionately less likely to pass. The most dire unintended consequence of failing an exam is dropping out of school. This paper explores the consequences of failing an exam for a group of HS students with the highest dropout rate of any demographic group and who are one of the fastest growing student population, English language learners (ELLs).

ELLs make up roughly a tenth of all U.S. students and about 15% of students in urban schools. NYC, the setting for this paper, has the second largest concentration of ELLs after Los Angeles. The 4-year and 5-year HS dropout rates for ELLs in 2015 were 22% and 33%, respectively, about 15 to 28 percentage points higher than the rate for students who are White, Black, Hispanic, Asian, economically-disadvantaged, or have a disability. Compared to HS graduates, dropouts and GED recipients have poorer earnings, hours of work, employment, job tenure, and health. Therefore, high dropout rates present an educational and social policy concern. The evidence on the impact of graduation exams is, at best, heterogeneous, with little evidence on impacts for ELLs. An important takeaway from the existing literature is that graduation exams can create both positive and negative incentives for students to either work harder and perform better or get discouraged and drop out. Disentangling these mechanisms, however, is empirically challenging and has never been done for ELLs. This paper addresses the following research questions:

1.  Does failing a graduation exam increase ELL’s likelihood of dropping out of HS?
2. Is the change in the likelihood of dropping out driven by students who failed and were discouraged or students who passed and were motivated?
3. Is there a long-term effect of failing an exam on the probability of college enrollment?

I use detailed, student-level, longitudinal data on all NYC public school students from the NYC Department of Education matched with postsecondary data from the National Student Clearinghouse (2005-2015). I will exploit a sharp regression discontinuity design to estimate the impact of failing an exam on HS dropout and college enrollment. Identification of causal effects is achieved by comparing students who are similar but on different sides of a predetermined test score cutoff. Additionally, I will use a policy change that effectively raised exam standards in NY in 2008 to disentangle any effects into encouraging versus discouraging factors. Findings from this paper will shed light on the mixed results found in the literature and present new evidence on treatment effects and mechanisms by which exams influence ELLs.