[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

Re: st: Question on how to analyze multiple imputation data using svylogit

From   "Austin Nichols" <[email protected]>
To   [email protected]
Subject   Re: st: Question on how to analyze multiple imputation data using svylogit
Date   Thu, 24 Jan 2008 11:02:48 -0500

Note that the program was just updated last week via the SJ
( The help files at e.g.

view "";
view "";
view "";

show that -svylogit- (I will guess you are using Stata 8.2 from the
mention of this outdated command) is not one of the supported

net describe mice, from(

shows this advice:

Report problems to the author  (who is _not_ on Statalist).
      Author: Patrick Royston, UK Medical Research Council
      Support: email [email protected]

So you may want to contact the author to see if svy commands are on
his near-term agenda, or if there is a work-around.

On Jan 24, 2008 10:31 AM, Malinda Kennedy <[email protected]> wrote:
> Dear group,
> I am a fairly new stata user, so I apologize for the simplicity of my question. I am trying to analyze a complex survey using the svylogit command. However, the dataset is a multiple imputation dataset.  The end result I am looking for is an overall odds ratio that takes into account all 5 multiple imputation datasets and the survey design. (I also will need to be able to use the subpop option as well.)
> What I have done so far:
> I imported the 5 datasets and then used "miset" to load the 5 datasets.
> I think if I just use "svylogit" now that stata will only use the first dataset-correct?
> So I tried using "mifit: svylogit ....." but I got the following message: "combining estimates is not possible: no variation between datasets" -
> I tried using "micombine svylogit....." and got a message saying it is an invalid command
> I tried using "mido svylogit...." and it looks like it runs for each dataset, but does not give an overall estimate.
> I have looked online and in the stata listserv archives and could not find an answer to my question. I am hoping someone can help me with what I assume is probably fairly easy if I can just find the code.
*   For searches and help try:

© Copyright 1996–2024 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index