UC San Diego

My dissertation is a collection of essays on effects of educational reforms on student outcomes. Chapter 1 investigates the effects of the California High School Exit Exam on student achievement along the entire student distribution. I report three main findings. First, the exit exam lowers graduation rates among students in the left tail of the achievement distribution. Second, the initial, more difficult version of the exam is associated with positive effects on achievement across the student distribution, with the exclusion of the lowest performing students. Finally, the current version of the exam is associated with negative effects on student test scores and grades among the low achieving students, and with increased achievement in the middle of the distribution. Chapter 2 analyzes the effects of initial failure on the California High School Exit Exam on student outcomes using the Regression Discontinuity framework. For students close to the passing cutoff, the initial failure does not affect student achievement while in high school. However, failing either component of the test increases the likelihood of enrollment in remedial classes the following year. An analysis of postsecondary outcomes shows that conditional on graduating from high school; initial CAHSEE failure has no effect on 2-year college enrollment or the likelihood of transfer to a 4-year university. Chapter 2 shows that the California High School Exit Exam does not pose a significant hindrance to student achievement for students close to the passing cutoff. Combining a unique annual longitudinal dataset on learning outcomes with a vertically-scaled item-response theory model, Chapter 3 presents the first estimates of learning trajectories over time for a given cohort of students in a developing country. We provide the first out of sample validation of a vertically-scaled IRT model in a low-income setting. We document increasing inequalities in learning outcomes as students go through primary school. We document significantly increasing variance in learning over time and increasing gap between the mean learning levels and the standard specified by the syllabus over time. Our results suggest that education systems in developing countries have not effectively made the transition from serving a screening function to providing human capital to all students