Hi all,
I am using svyglm and svyttest to weight my outcome with propensity scores
as per:
Ridgeway & colleagues (2013) "Toolkit for Weighting and Analysis of
Nonequivalent Groups: A tutorial for the twang package"
So after:
>glm1 <- svyglm(X ~ Y, design=design.ps)
>summary(glm1)
or
>svyttest(X ~ Y, design=design.ps)
(where X is my dependent variable and Y, my factor), I obtain my means
using:
>svyby(~X, ~ Y, svymean, design=design.ps, na.rm=TRUE)
I also determine "doubly robust" estimates to control for covariates that
remain unbalanced after propensity score weighting using:
>glm1 <- svyglm(X ~ Y + a + b + c, design=design.ps)
>summary(glm1)
or
>svyttest(X ~ Y + a + b + c, design=design.ps)
(where a, b and c are my covariates).
My question is how to obtain adjusted means and standard errors relating to
my new model. The use of:
>svyby(X ~ Y + a + b + c, svymean, design=design.ps)
is extremely inefficient (and I am yet to obtain an estimate).
Many thanks,
Andrew Kemp
University of Sao Paulo
[[alternative HTML version deleted]]
______________________________________________
R-help@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
