Using Hierarchical Linear Growth Modeling to Identify Longitudinally Outperforming School Districts in the United States, 2009–2013

Abstract

ABSTRACT School District Effectiveness Research (SDER) works to identify unusually effective organizations for in-depth qualitative study that significantly outperform peer districts over multiple years. Yet, critiques of SDER include a focus on cross-sectional data, individual states, and percent proficient metrics. Here, we analyze the Stanford Education Data Archive, including data on n = 10,825 districts 2009–2013, from 47 states and DC, with average student achievement equated to the National Assessment of Educational Progress (NAEP). To replicate previous state-level SDER, we use a two-level hierarchical linear growth model, examining district growth over time, including an AR1 autoregressive parameter. We identify 536 outperforming districts nationally.