Association for Public Policy Analysis & Management

Research has documented that teacher quality is highly variable, even within a single school, whether measured by teachers’ ability to raise their students’ learning over the course of a school year (Hanushek and Rivkin 2010; Chetty, Friedman and Rockoff, 2014) or by key measures of classroom practice, such as the use of class time and ability to keep students engaged (Bruns and Luque, 2015). The ability of schools and school systems to offer teachers performance feedback, incentives and appropriate supports to improve their performance, crucially depends on valid measures of their effectiveness.

The Classroom Assessment Scoring System (CLASS) instrument has been validated in over 2,000 US schools, with evidence that teachers’ ratings on the instrument predict their students’ value-added learning gains, and is beginning to be used in developing countries like Ecuador and Chile. CLASS is attractive because it creates a comprehensive measure of teacher skill across 3 key domains: emotional support; classroom management; and instructional support. To date, CLASS has mainly been used in research settings because it requires highly trained professionals, making it very costly.

The Stallings Classroom Snapshot instrument focuses on more limited dimensions of teacher practice –classroom management and student engagement– and can be applied reliably by observers with less than one week of training, which makes it useful for larger scale studies. Research in 6 Latin American countries has established correlations with student learning at the school level, but these applications did not test correlations at the classroom level.

This is the first research study to compare the CLASS and Stallings instruments. The study aims to answer:

-       How consistent are the assessments of teacher quality using the two instruments?

-       After identifying the sub-dimensions of CLASS most similar to the constructs measured by Stallings: How strong are the correlations between each instrument’s measures on the dimensions and student learning outcomes?

-       Is a teacher’s ability to manage the classroom a proxy for the broader assessment of quality?

Hypothesis: We expect Stallings to be correlated to the classroom management sub-dimension of CLASS, but not necessarily to the overall CLASS score. We expect Stallings scores to be correlated with student learning, but we also expect the correlation to be weaker than CLASS scores with student learning.

DATA:

- 102 high quality classroom videos of 51 seventh grade math teachers in Chile. Teachers were filmed at two different points in the school year.

- SEPA student learning tests conducted at the beginning and end of the year in the same 51 classrooms.

The findings of this study are valuable for policymakers in search of instruments to evaluate and improve teacher quality. Researchers conducting impact evaluations of interventions aimed at raising teacher effectiveness can also benefit from the evidence this study will generate. As the Stallings instrument has lower costs and requires less sophisticated enumerator training, if validated, it could be a useful substitute for CLASS in specific instances. The study will indicate the value for the two different instruments, and the most appropriate contexts for their use.