This paper examines the relationship between teacher qualifications and student achievement, using administrative data from North Carolina public schools. It documents the tendency for more qualified teachers to be matched with more advantaged students, which can introduce bias into estimates of teacher effectiveness. To minimize this bias, the authors restrict their analysis to schools that assign students to classrooms in a manner statistically indistinguishable from random assignment. They find significant returns to teacher experience in both math and reading, and to licensure test scores in math achievement. They also find that the returns in math are greater for socioeconomically advantaged students, suggesting that the observed form of teacher-student matching persists in equilibrium. The paper also explores the effects of sorting processes on estimates of teacher effectiveness. It finds that teacher experience and licensure test scores are robust determinants of test scores for fifth grade students. The authors use a variety of strategies to counter the bias from sorting processes, including the addition of student-level control variables, the use of school fixed effects, and the use of a subsample of schools with balanced distributions of students across classrooms. They find that the bias from between-school sorting is large, while the bias from within-school sorting is more limited and may vary in sign across subsamples of schools. The paper also finds that the effects of teacher qualifications on student achievement vary by type of student. For example, the returns to teacher experience are significantly larger for students not receiving subsidized lunches, indicating that more advantaged students benefit more from experienced teachers. The authors conclude that the observed pattern of teacher-student matching is consistent with the goals of teachers, parents, and administrators, and that it may explain why the alternative pattern of negative matching is not the empirical norm. The paper highlights the importance of using detailed administrative data to disentangle omitted variable bias from true causal effects and to estimate the effects of teacher qualifications on student achievement.This paper examines the relationship between teacher qualifications and student achievement, using administrative data from North Carolina public schools. It documents the tendency for more qualified teachers to be matched with more advantaged students, which can introduce bias into estimates of teacher effectiveness. To minimize this bias, the authors restrict their analysis to schools that assign students to classrooms in a manner statistically indistinguishable from random assignment. They find significant returns to teacher experience in both math and reading, and to licensure test scores in math achievement. They also find that the returns in math are greater for socioeconomically advantaged students, suggesting that the observed form of teacher-student matching persists in equilibrium. The paper also explores the effects of sorting processes on estimates of teacher effectiveness. It finds that teacher experience and licensure test scores are robust determinants of test scores for fifth grade students. The authors use a variety of strategies to counter the bias from sorting processes, including the addition of student-level control variables, the use of school fixed effects, and the use of a subsample of schools with balanced distributions of students across classrooms. They find that the bias from between-school sorting is large, while the bias from within-school sorting is more limited and may vary in sign across subsamples of schools. The paper also finds that the effects of teacher qualifications on student achievement vary by type of student. For example, the returns to teacher experience are significantly larger for students not receiving subsidized lunches, indicating that more advantaged students benefit more from experienced teachers. The authors conclude that the observed pattern of teacher-student matching is consistent with the goals of teachers, parents, and administrators, and that it may explain why the alternative pattern of negative matching is not the empirical norm. The paper highlights the importance of using detailed administrative data to disentangle omitted variable bias from true causal effects and to estimate the effects of teacher qualifications on student achievement.