JSM 2004 Online Program

Abstract #302158

This is the preliminary program for the 2004 Joint Statistical Meetings in Toronto, Canada. Currently included in this program is the "technical" program, schedule of invited, topic contributed, regular contributed and poster sessions; Continuing Education courses (August 7-10, 2004); and Committee and Business Meetings. This on-line program will be updated frequently to reflect the most current revisions.

To View the Program:
You may choose to view all activities of the program or just parts of it at any one time. All activities are arranged by date and time.

The views expressed here are those of the individual authors
and not necessarily those of the ASA or its board, officers, or staff.

Back to main JSM 2004 Program page

Activity Number:	311
Type:	Contributed
Date/Time:	Wednesday, August 11, 2004 : 8:30 AM to 10:20 AM
Sponsor:	General Methodology
Abstract - #302158
Title:	Testing for Latent Unit-by-treatment Interaction in Clustered Randomized Trials with Binary Outcomes
Author(s):	Ed Mascha*+
Companies:	Cleveland Clinic Foundation
Address:	9500 Euclid Ave, Cleveland, OH, 44026,
Keywords:	potential outcomes ; unit-by-treatment interaction
Abstract:	Individual treatment effect heterogeneity, also called unit-by-treatment interaction, is a quantification of the proportion of the population that responds differently to one treatment than to the other. It is quite distinct from the usual treatment-by-covariate interaction that is based on differing marginal treatment effects across levels of a covariate. With a given covariate, there may exist both types of interactions, one or the other type, or neither. Unit-by-treatment interaction may well vary over levels of a covariate or covariate patterns, but it is independent of the marginal treatment-by-covariate interaction with any covariate. It may be conceptualized and studied regardless of whether there are covariates related to it. We develop a method to test this type of interaction in parallel group randomized studies where we only observed 1 of the 2 binary outcomes of interest for each individual. The method depends on natural clustering in the data. We use a de-clustering algorithm to estimate the relationship between variabels known to provide information about the interaction of interest. Rejecting the null hypothesis in our test implies non-independence of

The address information is for the authors that have a + after their name.
Authors who are presenting talks have a * after their name.

Back to the full JSM 2004 program