Panel Paper: Getting Better? Estimating the Impact of Participation in a Multiple Measure Teacher Evaluation System on Teacher Quality

Thursday, November 3, 2016 : 1:55 PM
Columbia 3 (Washington Hilton)

*Names in bold indicate Presenter

Katharine O. Strunk, Edward Cremata and Julie A. Marsh, University of Southern California


The ways in which teachers are evaluated has changed drastically in the last decade. States and districts across the country have implemented multiple measure teacher evaluations systems (MMTES) incorporating classroom observations, value-added measures (VAMs) of teacher contributions to student achievement, and other factors (e.g., stakeholder feedback, contributions to the community) in an effort to improve the rigor with which they evaluate teacher performance, as well as to provide targeted supports to improve teacher practice (Doherty & Jacobs, 2015). By 2015, 43 states required school districts to evaluate teachers using MMTES that incorporated measures of teachers’ contributions to student achievement, and 23 states required that measures of teacher quality based on these evaluations must inform tenure decisions. These numbers have increased from 15 and zero, respectively, in 2009 (Doherty & Jacobs, 2015). Given the recent nature of these policy shifts, evidence is just beginning to emerge about the efficacy of new MMTES in terms of their impacts on teacher quality, teacher mobility and student achievement. The extant literature that addresses these questions finds that teacher evaluation systems incorporating classroom observations and feedback on teacher practice, in addition to measures of teachers’ contributions to student achievement growth, can generate sizeable improvements in student achievement (Taylor and Tyler, 2012; Steinberg and Sartain, 2015) and teacher performance (Dee and Wyckoff 2013).

In this paper we examine the Los Angeles Unified School District’s (LAUSDs) new teacher evaluation system, the Educator Development and Support: Teachers (EDST), and its impact on subsequent teacher performance (measured using teachers’ VAMs). The EDST is based primarily on principals’ observations of teachers’ practice (teachers are typically observed twice a year) and assessments of teachers’ contributions to professional practice. Teachers are assigned to be evaluated at five-year intervals, as well as if they are new to the district or to their school, or if they were given a prior warning regarding performance or conduct.

We utilize a panel of teacher-level administrative data that includes teacher characteristics, measures of teacher quality (VAMs) and indicators of teachers’ EDST participation and their EDST performance (a continuous score ranging from 1-4 based on their observed practice and a final summative rating of “meets” or “does not meet” standards). We use linear regression and teacher fixed effects models to examine whether teacher performance increased as a result of participation in the EDST during its first two years of scale-up (2013-14 and 2014-15). We also assess whether the relationship between EDST participation and VAMs differs according to teachers’ average observational rating and final summative evaluation outcome. Characteristics affecting the relationship between EDST participation and teacher quality are examined as well, including school, principal, and teacher characteristics, as well as the proportion of EDST-participating teachers within each school.

Preliminary evidence suggests that participation in the EDST improves teacher performance, especially for teachers who are evaluated as a result of previous poor performance. In addition, teachers who saw the greatest increases in their observation-based scores over the course of the evaluation year saw greater gains in effectiveness.