|
Activity Number:
|
227
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, August 10, 2004 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Bayesian Statistical Science
|
| Abstract - #302033 |
|
Title:
|
A Bayesian Cross-classified Model for Estimating Teacher Effects on Student Achievement
|
|
Author(s):
|
J. R. Lockwood*+ and Daniel F. McCaffrey and Louis T. Mariano
|
|
Companies:
|
RAND Corporation and RAND Corporation and RAND Corporation
|
|
Address:
|
201 N. Craig St. Suite 202, Pittsburgh, PA, 15213,
|
|
Keywords:
|
value-added modeling ; teacher effects ; student achievement ; cross-classified model
|
|
Abstract:
|
In the education research and policy communities there is increased interest in using "value-added models," relying on longitudinal student-level test score data, to isolate the contributions of teachers to student academic achievement. The fact that students are linked to different teachers as they progress through grades imparts a cross-classified structure to the data that poses both substantive and computational challenges. We introduce a general multivariate, longitudinal model for student outcomes that incorporates these complex grouping structures, subsumes the principal existing modeling approaches, and explicitly allows for the estimation of the long-term effects of past teachers on student outcomes in future years. We present a Bayesian formulation of the model and discuss how the Bayesian framework affords numerous advantages relative to classical mixed-model approaches to value-added modeling. We demonstrate the model and the array of inferences about teacher effects that it facilitates using actual student achievement data.
|
