Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: st: regression diagnostics with complex survey data

From   Nick Cox <[email protected]>
To   "[email protected]" <[email protected]>
Subject   Re: st: regression diagnostics with complex survey data
Date   Wed, 3 Aug 2011 20:49:09 +0100

The Statalist FAQ has long carried the advice to flag if you are not using the latest version, now Stata 12, advice which you seemingly did not heed.

As I am away from base, I cannot identify what choices are available to you, but someone else may be able to look at Stata 9 and see what is available to you.

On 3 Aug 2011, at 20:34, Hitesh Chandwani <[email protected]> wrote:


I think I should explain myself more completely. I use Stata 9. So
"svy: glm" does not work. I have to do a regular glm and then do
"suest ., svy"..this works. But I can't figure out how to run
postestimation diagnostics on this analysis. Any postestimation
commands will use results from the regular GLM. I am not able to
figure out how to run postestimation while taking into account the
survey design.

Any ideas?


On Thu, Jul 28, 2011 at 9:31 AM, Nick Cox <[email protected]> wrote:
. help glm postestimation

mentions several commands with application to -svy-, so "none" is not correct.


On Thu, Jul 28, 2011 at 2:26 PM, Hitesh Chandwani
<[email protected]> wrote:

I'm posting my query as a response to this thread because it is a related topic.

I am also using survey data but my dependent variable is cost and is
positively skewed. I'm using a GLM with gamma distribution and log
link. Are there any post-estimation procedures that I could run? I
found a lot of post-estimation commands for linear regression but none
for GLMs.

*   For searches and help try:

© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index