The Econometrics and Economics of Education

Abstract

This thesis contains three essays spanning the fields of econometrics and the economics of education. The first is a methodological essay wherein I propose a solution to the test score measurement problem. Test scores that measure the same skill or trait are often scaled differently. I propose a statistical methodology to express test scores in a standard format to make them comparable across different tests. While other methods to standardize scores exist, the method I develop avoids several statistical pitfalls that commonly befall other procedures. The second chapter is an applied paper examining the question of whether academic achievement increases when students are the same race as their teacher. I pay particular attention to the effects of same-race teachers on academic achievement in both the short and medium run. Using a model that takes into account past school inputs, I find positive effects of racial matching on student achievement; moreover, these benefits persist with time. The third is a methodological piece that develops a technique to answer the question of whether one can be confident that a statistical estimate of something such as a mean or a regression parameter estimate lies between two points. The procedure has desirable statistical properties, and is ideally suited for use when testing for the existence of zero or near-zero effects.