Using Value-Added Assessment for Personnel Decisions: How Omitted Variables and Model Specification Influence Teachers’ Outcomes

• In Event: Policy Issues Related to Teacher Quality

Abstract

Value-added models (VAM) provide estimates of gains in student achievement that can be ascribed to specific teachers or schools. Most researchers are confident that VAM can be used to draw attention to teachers that may be underperforming and could benefit from additional assistance. However, they caution educators about using such models as the only consideration for high-stakes outcomes. This paper considers the impact of omitted variables on teachers’ value-added estimates, and whether commonly used single equation or two-stage estimates are preferable when possibly important covariates are not available. The findings indicate that these modeling choices can influence outcomes for individual teachers, particularly those in the tails of the performance distribution who are most likely to be targeted by high-stakes policies.

Authors

• Christine Mokher, CNA

• Linda Cavalluzzo, CNA Education

• Dale Ballou, Vanderbilt University