Panel Paper: Do Students Fairly Assess Teacher Effectiveness When It Matters?

Thursday, November 8, 2012 : 1:15 PM
Salon B (Radisson Plaza Lord Baltimore Hotel)

*Names in bold indicate Presenter

Richard Bowman, Albuquerque Public Schools and Sade Bonilla, Stanford University

The push for higher teacher quality through better evaluation methods has gained strength over the last several years. The U.S. Department of Education’s widely publicized Race To The Top competition and the number of states participating in the No Child Left Behind waiver process are the latest evidence of this force. Questions about the adequacy of an evaluation system relying solely on student test scores or even gains in student test scores have been raised by stakeholders, researchers, and policymakers. In response, many states and districts have begun piloting “multiple measures” for teacher evaluation. Student surveys, such as the Tripod Survey, are one measure that has received favorable reception by policymakers and researchers. Research by the Measures of Effective Teaching (MET) Project has suggested that student surveys may be able to shed light on teacher effectiveness. However, individuals’ survey responses may differ when stakes are attached to an outcome.  Nevertheless, to date, we have little information on whether teenagers will change their answers if they know that their answers have an effect on how much their teacher compensated based on the survey outcome.

This paper uses a within-classroom experimental design to show whether and how middle school and high school students respond differently when they know that their teacher will be paid based on their answers. This study includes four large urban public middle and high schools participating in a “multiple measures” teacher evaluation and compensation pilot. Students’ responses on a survey similar to the one used by the MET Project are linked to students’ academic and demographic background and teachers’ value-added results. Preliminary results from the first wave of approximately 6000 surveys suggest that students’ answers do not significantly change in their average level, teacher rankings, or correlations with teacher value-added if they know their teacher will get paid based on their answers, however this result may be moderated by student academic background and demographic characteristics.