Essays on human capital and economics of education
MetadataShow full item record
Since the publication of the Coleman Report in 1966, economists have contributed significantly to the body of research focused on understanding how school systems work. By approaching the development of students' human capital as a production process in which school resources, teachers, etc., are the inputs, we can utilize the rich empirical and theoretical tool kit developed to understand the production process of firms. However, recent studies show that there are more factors that have not been studied yet. Moreover, not all already known factors have studied thoroughly. My dissertation studies the following questions in an education production context: First, does alleviating adolescents' sleep deprivation improve academic performance? Second, do peer effects affect student performance even across grade? Third, is there a benefit to taking more tests? First chapter examines how the delayed school start time affects students’ sleep schedule, their time use on other activities as well as academic performance. I exploit a natural experiment that delayed school start times in a large province in South Korea using both the Korean Time Use Survey and school administrative data in a difference-indifferences framework. I find that teenagers and middle school students slept an additional 30 minutes per school night by spending less time studying independently. Despite the reduced study time, I find that delaying the school day in middle school start time by forty minutes increased math test scores by 0.034 standard deviations and reading test scores by 0.022 standard deviations. This confirms that reduced sleep deprivation is the main mechanism of improved academic performance. The second chapter examines how peer effects work within high school. Most studies on secondary education peer effects have focused on peer effects within a classroom or a grade, but students’ peer networks can be much larger, even across grade. Thus, I examine, in work with Jungmin Lee, whether students do better when the students in other grades are better. We find no effect of having more high-achieving students in another grade on your own performance, whereas we do find that having more low-achieving students has a negative impact on your performance. The third chapter examines the impact of eliminating two exams in one year on later test scores by exploiting a phased-in policy in Korea in which students are exempt from second-semester exams in their first year in middle school with a difference-indifferences strategy. I find that taking two tests less during the last school year of middle school decreased math test scores by 0.053 standard deviations, but does not have an impact on reading test scores. This negative impact on math test scores remains relatively constant for at least one academic year. This means more tests can incentivize students to spend additional time studying and in learning more knowledge or test taking skills. However, it is notable that less tests do not have a harmful effect on reading, but on math. This suggests the number of tests has an impact on students’ academic performance through learning knowledge than test taking skills since math requires more prior knowledge and is especially learned in class. Moreover, it is less likely that required test taking skills are different based on subject as most portion of the tests in the data is multiple choice questions.