Methodological Concerns About the Education Value-Added Assessment System (EVAAS): Validity, Reliability, and Bias

• Main Authors: Audrey Amrein-Beardsley, Tray Geiger
• Format: Article
• Language: English
• Published: SAGE Publishing 2020-05-01
• Series: SAGE Open
• Online Access: https://doi.org/10.1177/2158244020922224

The Education Value-Added Assessment System (EVAAS), the value-added model (VAM) sold by the international business analytics software company SAS Institute Inc., is advertised as offering “precise, reliable and unbiased results that go far beyond what other simplistic [value-added] models found in the market today can provide.” In this study, we investigated these claims, as well as others pertaining to the validity or truthfulness of model output, by conducting analyses on more than 1,700 teachers’ EVAAS results (i.e., actual EVAAS output to which no other external scholars have had access prior) from the Houston Independent School District (HISD). We found the EVAAS to perform, overall, in line with other VAMs in terms of validity and reliability, although it yielded possibly more biased value-added estimates than other VAMs due to differences in teacher’s EVAAS scores based on school-level student composition factors.