Principal Quality, and the Influence of Principals on Teacher Quality: Evidence from Panel Data

Through their control of, and influence over, multiple aspects of a school, principals are integral to improving student outcomes. In fact, they are considered one of the most important within-school factors, second only to teachers, in their effect on student experiences and success. As a result, when principals leave schools, their departure disrupts the functioning of the school, negatively affecting student and teacher outcomes.

Because principals rarely instruct students in a classroom setting, decades of research on principals theorize that much of principals’ influence on students is through other individuals and factors, especially teachers. It is argued that principals shape school structures that help teachers improve their practice. Through these practices as well as the differential hiring and retention of teachers, they also influence the composition of the teaching workforce. These multiple avenues serve as examples of how principals can affect student outcomes through the quality of the teaching workforce at their schools.

Due to a lack of large-scale data collection, researchers have only recently begun to systematically test whether and how principals use the aforementioned strategies to influence teacher quality, and how their ability to do as much varies by their measured quality. Very little work in the education leadership field, therefore, has used rigorous quantitative methods and leveraged large-scale panel data to understand the various aspects of the principalship. To add to this nascent area of research, my dissertation provides insights into each of following topics: principal effects on teacher quality and turnover, different measures of principal quality, and the relationship between principal quality and turnover.

Chapter 1 of this dissertation employs data from the District of Columbia Public Schools (DCPS) to measure principals’ contributions to improvements in teacher quality. DCPS arguably has the most well-developed and well-implemented teacher evaluation system in the U.S., IMPACT. I use teachers’ classroom observation scores as a measure of teacher quality. I find substantial variation across principals in their contributions to changes in teachers’ classroom observations scores. Results also show that principals make meaningful contributions to these scores.

Not only does DCPS have a valid and reliable teacher evaluation system, it also has a long-standing principal evaluation system (SL-IMPACT), a major component of which is intended to account for the quality of teaching. Using principal performance on this particular aspect of SL-IMPACT, I examine the extent to which DCPS’s measure of principal quality correlates with principals’ contributions to teacher observation scores. I find that DCPS’s measures as well as other, externally evaluated assessments of principal effectiveness, share small to moderate correlations with estimates of principal contributions to teachers’ classroom practice.

As discussed previously, principals can improve the teaching workforce at their schools through improving the quality of current teachers and/or affecting the composition of the teaching workforce. While Chapter 1 explores the former, Chapter 2 focuses on the latter. Specifically, Chapter 2 examines the effect of principals on teacher attrition. Using a 40-year panel of all public school teachers and principals in New York State, I (in collaboration with David Matsa and Amalia Miller) explore how female principals affect rates of teacher turnover – an important determinant of school quality. We find that male teachers are more likely to leave their schools when they work under female principals than under male principals. In contrast, we find no such effects for female teachers.

Chapter 3 returns to an idea that I touched upon in my discussion of Chapter 1: evaluating principal quality. I (along with Luke C. Miller and Dan Player) use teachers’ perceptions of their principals as a way to operationalize principal quality. We estimate the relationship between principal quality and principal turnover at the local level in New York City (NYC) as well as at the national level using the National Center for Education Statistics’ (NCES) Schools and Staffing Survey (SASS) data. Results show that the higher teachers rate their principal’s quality, across both contexts, the less likely the principal is to turn over.

These three chapters address underexplored areas of the principal literature. By using large-scale panel datasets and rigorous quantitative methods, I attempt to answer questions relating to principal quality and turnover, as well as principal effects on teacher quality. I also offer policy implications relating to each of my research questions, and future directions for research.