Organizational stability and school performance

Despite decades of policy innovation aimed at improving school performance, the number of public schools defined as low performing in the U.S. continues to grow. Yet, most explanations of low performance do not consider the fact that many of the country’s lowest performing schools share high rates of turnover among both staff and students, or organizational instability. The purpose of this dissertation was to develop the theoretical underpinnings of both the concept of organizational stability and its relationship with school contextual factors and performance, and to assess the relationship quantitatively. I hypothesized that teacher turnover, principal turnover, and student mobility partially mediate the impact of a school’s socioeconomic context on its academic performance. In order to test the proposed partial mediation model, I conducted a quantitative analysis using path analysis and data from North Carolina public schools. I constructed several samples, including one that included all schools, and five others that focused in on high instability, high poverty, and urban schools. The results of the analyses depended in large part on the type of school under investigation. Specifically, I found that the relationships between the context variables varied according to the sample being examined. Similarly, the presence of direct and mediating effects between the organizational stability variables was contingent on the kind of school. The results of these analyses support previous findings and contribute a new understanding of the role of instability in helping to explain low school performance. This dissertation engages the ongoing debate about the effects of teacher and principal turnover on school performance, suggesting that both do indeed have a deleterious effect on performance. Finally, I make several methodological contributions to the literature on both turnover and school performance by utilizing path analysis, which allows for the prioritization of effects and the testing of indirect effects.