Ben Domingue wrote:
>
> I'm not quite sure what you mean. If all you need is propensity
> scores to run an IPW analysis, the fitted values should work. Having
> many binary covariates shouldn't be a problem, the whole point of the
> propensity score is boiling down many dimensions to a single one.
> I use matchit() for my psm needs, but it may not be so useful for IPW.
>
> Ben
>
>
>
Yes, for now that is all i need. I just wanted to know if fitted.values is
all I need for IPW, because i was not sure how they were estimated by R. I
know the pscore is boiling it down to one scalar, just said that in reply to
Frank E Harrell advocating for covariate adjustment, because i thought he
referred to all the covariates. Anyway, thx Ben, this helped a lot. Just
wanted to use pscores apart from all the matching algorithm and so forth.
So fitted.values are essentially the probabilities for inclusion ?
--
View this message in context:
http://www.nabble.com/propensity-score-adjustment-using-R-tp19555722p19558363.html
Sent from the R help mailing list archive at Nabble.com.
______________________________________________
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.
