Estimating the Standardized Average Treatment Effect with Missing Outcome Data

USC Quantitative Speaker Series (Spring 2023)

Date: March 9, 2023

Speaker: Samantha Anderson, Ph.D.

Assistant Professor
Department of Psychology - Quantitative
Arizona State University

Video Recording (requires sign in using your USC NetID)

Abstract

In this talk, I will present research assessing estimation of the standardized average treatment effect (sATE) in randomized pretest posttest studies when outcome data are missing. Standardized effects are increasingly reported in treatment studies, to facilitate interpretation, future meta-analysis, and study comparison. However, when outcome data are missing, achieving an unbiased, accurate estimate of the sATE can be challenging, given that the sATE is a ratio of a (sometimes adjusted) mean difference to a (within-group) standard deviation. The simulation study I will describe compares missing data strategies, specifications, and modeling choices in terms of bias and accuracy. Populations with homogeneous and heterogeneous variances are considered, as well as multiple different missingness patterns relevant to these types of studies. The results show that although the sATE can be estimated with little bias using either maximum likelihood or multiple imputation, particular attention should be paid to the model and variance estimator, especially at smaller sample sizes. In addition to presenting simulation results, I will include demonstration of how and why certain choices lead to bias when estimating the sATE.