EDM
Improved Power in a Reanalysis of an ASSISTments Efficacy Trial by Leveraging Auxiliary Data
Thu Jul 2, 2:40 PM–2:55 PM · North 209
Educational Data Mining Methods Psychometrics & Educational Measurement Intelligent Tutoring Systems
★ Notable speakers
Adam Sales
★
— Causal inference in educational data mining; combining machine learning with randomized trial analysis; principal stratification in ITS
Mingyu Feng
★
— Large-scale randomized controlled trials evaluating intelligent tutoring systems (ASSISTments, MathSpring); IES-funded efficacy research on ed-tech
Johann Gagnon-Bartsch
★
— Causal inference; statistical methods for unobserved confounders; RUV (removing unwanted variation)
A reanalysis of an ASSISTments efficacy trial showing that incorporating auxiliary data increases statistical power for detecting learning gains.