Predicting Gain Scores with Hierarchical Linear Models: A Value-Added Approach to Measure Teacher Effectiveness

Abstract

In this article, we predicted students’ mathematics gain scores employing two-level hierarchical linear models (HLM) through value-added approach using data from one of the largest urban school districts in the United States of America. Effects of teacher quality or teacher effectiveness, characterized by teacher’s certification in mathematics content area and teacher experience, were measured on students’ gain scores. The results showed significant impact on mathematics gain scores due to teacher’s content certification and teacher experience at teacher level and pretest scores as well as free and reduced lunch status at student level including cross-level interaction effects of teacher content certification with student level predictors. We also reported proportions variance explained and d-type effect sizes for teacher level models in order to measure teacher effectiveness.