Describing the Malleability of Teaching Practice

• In Event: 63.018 - Promising Scholarship in Education: Dissertation Fellows and Their Research

Abstract

Despite considerable policy interest in improving teacher performance, we know very little about which dimensions of classroom practice are most malleable over time. Researchers have documented effectiveness gains writ large during the course of the teaching career, but the lack of longitudinal data on classroom practice has prevented more granular exploration of these changes. Understanding the malleability of teaching practice has important implications for hiring, evaluation, and professional development and could help districts target efforts in each area more effectively. By pairing rich observation data from the Measures of Effective Teaching (MET) study with a longer panel of observation data from the District of Columbia Public Schools (DCPS) teacher evaluation system I hope to contribute an initial description of the malleability of teaching practice in these unique contexts.

Research Questions
1. Which dimensions of classroom practice exhibit the greatest change over time?
2. How does the rate of growth for dimensions of practice relate to the level of the initial score?
3. Is the rate of change over different dimensions of practice predicted by teacher attributes such as experience or education?
4. Is the rate of change over different dimensions of practice predicted by school attributes such as poverty status or school-wide achievement?

This analysis will employ longitudinal classroom observation records to understand how teachers’ performance on various components of classroom practice varies over time and how this change is predicted by teacher and school attributes. The primary empirical model in the analysis will be a multilevel model for change. This approach has the advantage of providing flexibility in modeling growth trajectories in teacher practices while simultaneously accommodating nested error structures in the data. The nested structure also allows for varying time points and frequency of observation, which can be a troublesome challenge in modeling growth. There is a well-defined literature in quantitative psychology and statistics that also supports this method of analysis for understanding the predictors of individual change over time

Author

• Melinda K Adnot, University of Virginia