Wednesday, Mar. 15, 2017
Congratulations to Elizabeth Monroe, a Masters student in Applied Statistics, who will be presenting her poster at the 2017 Modern Modeling Methods Conference!
Poster Presentation Topic:
School district effectiveness research is a growing field in education research, but it lacks a method for identifying districts that significantly and longitudinally outperform peer districts from similar community and local contexts. In this study, we analyze the US population of public school districts over the five years 2009-2013. We built a two-level hierarchical linear growth model that nests time in districts and has parameters on both the intercept, in 2009, and change in slope over time, as well as including an AR1 autocorrelation specification. Using this model, we identify school districts that significantly outperform or underperform their predicted longitudinal student performance, given demographic and community covariates. The model’s covariates are variables district administrators cannot control, such as student demographics and state funding. Therefore, districts that outperform the model based on the residual analysis are significantly outperforming their peers, after controlling for such characteristics. The aim of this study is to present a method for selecting the most effective school districts in the United States, so that leaders of these districts may share how they have led their school districts to success. Several outperforming school districts are identified as potential sites for qualitative studies of district effectiveness.