Paper Summary
Share...

Direct link:

Robust Small Area Estimation for Generalization (Poster 20)

Fri, April 22, 11:30am to 1:00pm PDT (11:30am to 1:00pm PDT), San Diego Convention Center, Floor: Upper Level, Sails Pavillion

Abstract

Generalization studies in education are typically based on samples of 30 to 70 schools whose results are generalized to populations that are least 20 times larger. This small sample to population size ratio limits the precision of design-based estimators of population average treatment effects. Prior work has shown the potential of small area estimation methods to improve generalizations from small samples. However, small area estimators are model-based so their validity depends on the model assumptions. This study assesses the performance of a robust small area estimator when core model assumptions are violated. We compare its performance to that of a commonly used small area estimator and identify the conditions, if any, under which the robust small area estimator offers improvement.

Authors