[R] multiple imputed files

Anthony Damico ajdamico at gmail.com
Tue Jan 27 04:28:30 CET 2015


On Mon, Jan 26, 2015 at 2:13 PM, hnlki <annelies.hoebeeck at ugent.be> wrote:

> Thank you for your answers.  In fact I am using the HFCS dataset,



cool, so survey data.  asdfree.com is a good place for examples.  also, if
you can match any officially-published statistics, i would love to
collaborate on a post with you  :)


> but I need
> 3-SLS and SUR. Does the code work with system.fit as well?
>

i am not sure if those techniques have been implemented for survey data
with R.  check

http://r-survey.r-forge.r-project.org/survey/

you might be able to do it without the survey package, but donig so will
incorrectly reduce your variance calculations (and make you more confident
of a result than you should be), i believe


> My  5 imputed files are stacked together in one dataset. In order to use
> <imputationList> (mitools) I need several imputed datasets. Does that mean
> that I have to split my dataset first, in order to create the
> imputationList?
>
>
yes, follow any of the examples with "mitools" on the asdfree.com github
repository where i create five data.frame objects before making an
imputationList object

	[[alternative HTML version deleted]]



More information about the R-help mailing list