| This dissertation examines how different inputs into the education production function translate into academic outcomes. The input examined in Chapters 1 and 2 is schooling quantity, and the inputs examined in Chapter 3 are observable teaching assistant characteristics and teaching quality.;In Chapter 1, I utilize an original dataset on weather-related school closures in Virginia to study the impact of missed class days on performance on the Virginia Standards of Learning exams. Exogenous variation in weather-related school closures within school districts is used to identify the effect of unscheduled school closures on test performance. I find a generally negative and statistically significant relationship between school closures and test performance in reading at the elementary and middle school levels. For example, in fifth grade a missed school day is associated with a 1.1 point decline in math scores and a 0.35 point decline in reading scores (0.03 and 0.01 standard deviations, respectively). I also find evidence suggesting that the negative impact of a missed school day on test performance is greater for school districts with a higher proportion of economically disadvantaged students.;In Chapter 2, coauthored with Maria D. Fitzpatrick and David Grissmer, we exploit quasi-randomness in the timing of assessment dates in the Early Childhood Longitudinal Study - Kindergarten Class of 1998 to examine how time in school affects achievement in kindergarten. Conservative estimates suggest that children's math and reading scores increase by 0.6% and 0.5% of a standard deviation, respectively, with an additional day between fall and spring tests. The results are statistically significant and extremely robust to specification choice.;In Chapter 3, I utilize data on students in Principles of Microeconomics at the University of Virginia to examine the effect of teaching assistants on a variety of student outcomes. The structure of this course offers an ideal setting to examine the effects of teaching assistant characteristics on student performance because of the essentially random matching between teaching assistants and students. I find that teaching assistants influence student performance but observable teaching assistant characteristics like gender, nationality and experience do not predict student grades or continued study in economics. |
