Methodological Concerns About the Education Value-Added Assessment System

A recent article published in Educational Researcher by Audrey Amrein-Beardsley examines methodological issues in the Education Value-Added Assessment System (EVASS) on the eve of reauthorization accountability provisions in the No Child Left Behind Act (NCLB). Amrein-Beardsley argues that although EVASS is currently the most sophisticated value added model available – or least bad – there are still several important methodological concerns that threaten the validity of findings generated using this model. The author aruges these issues must be addressed before widespread adoption. Methodological concerns with EVASS include: 1) Missing and miscoded data, particularly over time; 2) regression to the mean; and 3) No control for student background variables. Additionally, Amrein-Beardsley argues there are insufficient validity studies, difficulties with user-friendliness of the model, and a lack of external reviews. Furthermore, the author reminds readers that it has yet to be demined whether high-stakes tests can be used to make valid inferences about student, knowledge, teacher effectiveness, or school/district effectiveness. As such, Amrein-Beardsley cautions policy makers, statisticians, and test developers against whole-sale endorsement of EVASS and highlights the need for additional research on the validity and methods of this model.

Key Words: academic achievement • accountability • achievement gains • educational legislation • evaluation methods • K-12 education • measurement techniques • teacher effectiveness