Association for Public Policy Analysis & Management

This study measures the effects of female school leadership on teacher turnover in K-12 education. Teacher turnover is an important driver of education productivity, but its causes are not well understood. Turnover lowers students’ academic achievement in classrooms that experience turnover but also in other classrooms in the school (Ronfeldt et al., 2013). Because the effects of turnover are larger in schools with greater proportions of minority and low-performing students (Ronfeldt et al., 2013), turnover contributes to inequality in educational outcomes. The cost savings from reducing turnover can also provide substantial value to school districts by reducing recruiting and training costs (Barnes, Crowe, & Schaefer, 2007).

We examine the relationship between principal gender and teacher retention using administrative data from the New York State Education Department on public schools from 1970 to 2010. The data include information on schools, principals, teachers, and students. The management literature finds that women lead organizations differently than men. Rather than adopting characteristically male behavior, successful female leaders in other settings adopt a distinctive, transformational style (Rosener 1990; Bass & Avolio 2006; Dezső & Ross 2012). For example, while men are more likely to assert themselves in a controlling manner, women tend to take greater account for the rights of others (Eagly and Carli 2007). Indeed, women business leaders are less likely to downsize their workforce in recessions (Matsa & Miller, 2013, 2014). In this study, we examine whether similar gender differences in leadership are also present in the school setting.

We also examine whether the relationship between principal gender and teacher retention varies with the gender of the teacher. Such a pattern mirrors relationships in the private sector, where female leaders are associated with improved career outcomes for female workers (e.g., Matsa & Miller, 2011; Kuzne & Miller, forthcoming).

We address these questions by estimating regression models for the effects of female principals on teacher retention, conditional on a range of school-level characteristics, including student race, percentage of free and reduced-price lunch eligible students, and average student achievement, as well as teacher and principal characteristics. This is similar in spirit to work by Grissom et al. (2012), who study data from the Department of Education’s 2003-2004 Schools and Staffing Survey, and find cross-sectional associations between principal gender and teacher satisfaction and turnover. However, our study differs from prior work by using panel data. Because we observe schools changing principals over time, we can control for unobserved heterogeneity across schools that may be correlated with principal gender by including a full set of school fixed effects.

After estimating the overall effects of female principals on turnover rates for male and female teachers, we also explore heterogeneous effects of female principals across different types of schools, such as elementary versus secondary (female principals are far more common in the former) and urban versus rural settings. Finally, we also examine if and how the effects of female principals have changed over the decades with the gender composition of the overall workforce.