[gov3009-l] VanderWeele on "Mediation and spillover effects in group-randomized trials"
mblackwell at iq.harvard.edu
Mon Feb 8 15:50:25 EST 2010
We hope you will join us this Wednesday, February 10th at the Applied
Statistics workshop when we will be happy to have Tyler VanderWeele
(Harvard School of Public Health). Details and an abstract are below.
A light lunch will be served. Thanks!
"Mediation and spillover effects in group-randomized trials: a case
study of the 4R's educational intervention."
Department of Epidemiology and Biostatistics
Harvard School of Public Health
February 10th, 2010, 12 noon
K354 CGIS Knafel (1737 Cambridge St)
In many group-randomized trials, questions of mediation are of
interest. In the education setting, often the treatment is at the
school level, the mediator is at the classroom level and the outcome
is at the individual level. The standard approach to mediation
analysis in the social sciences is subject to three important
shortcoming: (1) failure to account for possible treatment-mediator
interaction, (2) failure to account for mediator-outcome confounding
(even if the treatment is randomized the mediator is not and is thus
subject to selection-confounding issues) and (3) failure to account
for the spillover effects of one classroom on the outcomes of children
in other classroom. Ignoring any of these three issues can lead to
biases and misleading inferences. Using a counterfactual
conceptualization of direct, indirect and spillover effects, we
provide a framework that can accommodate all three of these issues.
We show that the total effect can be decomposed into what is defined
as a natural direct effect, a within-classroom mediated effect and a
spillover mediated effect. We give identification conditions for each
of the causal effects of interest and we provide theoretical results
on the sequences of ignoring "interference" or spillover effects when
they are in fact present. The methodology is applied to a
group-randomized trial of the 4R's educational intervention.
Institute for Quantitative Social Science
Department of Government
email: mblackwell at iq.harvard.edu
More information about the gov3009-l