Final Defense: Sarah Stephens
Tuesday, November 14, 2017, 11:45am

Sarah Stephens, UT-Austin

"Longitudinal Predictions Using Alternative Binning to Reduce Regression to the Mean"

Abstract: The analysis of standardized testing data must be done rapidly and accurately if it is to inform education policy. I have developed an alternative binning technique that reduces the regression to the mean in predictions of longitudinal testing data. Inspired by streamlines in fluid mechanics, alternatively binned streamlines accurately predict scores throughout primary and secondary school using only 2-3 years of data. Additionally, these streamlines can identify the effects of interventions, such as the Student Success Initiative. Alternatively binned streamlines could be used to bridge the gap between policy makers and researchers, allowing for more informed policies.

Location: RLM 9.222