How Well do Standards-based Teacher Evaluation Scores Identify High-quality Teachers? A Multilevel, Longitudinal Analysis of One District

Meeting Paper

States and districts functioning under the federal No Child Left Behind (NCLB) legislation and state standards-based reforms are being measured and potentially sanctioned by their ability or inability to staff every classroom with a highly qualified teacher, ensure that all students make adequate yearly progress (AYP), and submit a plan to address inequitable distributions of teacher quality. All of this emphasis on teacher quality, combined with a general agreement among education researchers with various backgrounds about the importance of teacher quality to student learning, has generated a need for states and districts to identify teachers who are capable of helping students learn to expected levels. One such means of identifying teachers that holds some promise is standards-based teacher evaluation systems. Previous research using a two-level hierarchical linear model has shown that they are valid, reliable means of identifying high-quality teachers (Milanowski & Kimball, 2005; Holtzapple, 2003).

This study improves upon previous studies by analyzing the effect of teachers with varying standards-based teacher evaluation scores in the context of a three-level model with controls for student-, teacher- and school-level characteristics. This model more accurately estimates the impact of higher or lower teacher evaluation scores on student achievement. It also accounts for the probability that school-level characteristics play a role in determining how effective teachers are in facilitating student learning, as well as the likelihood that teachers are not evenly distributed across schools, as the previous two-level model assumes. In addition, using three years of data, this study evaluates the stability of standards-based teacher evaluation scores, an important indicator of how well they measure teacher quality.

Download:
Publication date: 
January 2007