Value-Added Subscores for Instructional Feedback: A Comparison of Univariate and Multivariate Prediction Methods

• In Event: 65.079 - Roundtable Session 23
  In Roundtable Session: 65.079-15 - Issues in Value-Added Modeling

Abstract

Value-added scores are rarely calculated for formative purposes, but teachers in one school district have expressed interest in strand-level scores that could help them improve instruction. How might value-added subscores be estimated? Can they be augmented to increase their reliability and diagnostic potential? This paper applies multivariate generalizability theory to value-added subscores with real data. Results will help advance scholarship around value-added modeling as an assessment endeavor and provide guidance to practitioners seeking ways to make value-added scores more useful for teachers.

Authors

• Christopher T. Moore, Minneapolis Public Schools

• Daniel Glass, Minneapolis Public Schools