Bookmark and Share

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


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

st: RE: Using command svy glm to obtain risk ratios


From   "Hussein, Mustafa (Mustafa Hussien)" <[email protected]>
To   "[email protected]" <[email protected]>
Subject   st: RE: Using command svy glm to obtain risk ratios
Date   Mon, 17 Feb 2014 16:32:35 +0000

Hello,

I am not sure why your -svy- procedure won't give eform coefficients (RRs) with the log-binomial GLM (which you are trying to run here). A better alternative to log-binomial regression for estimating RRs (or prevalence rate ratios if cross-sectional data) is a modified Poisson (Poisson w robust SEs). The log-binomial is well-known to have issues with convergence and, as you've experienced, takes forever. A Poisson model is very easy to do using -glm- or -poisson-, and with pweighted-survey data, you've got to just add the eform option (or -, irr- with poisson).

Hope that helps
Mustafa    
________________________________________
From: [email protected] [[email protected]] on behalf of Agunwamba, Amenah A., Ph.D. [[email protected]]
Sent: Monday, February 17, 2014 9:47 AM
To: [email protected]
Subject: st: Using command svy glm to obtain risk ratios

Hello Stata Experts,

I'm attempting to obtain risk ratios using weighted survey data. I'm using a binary outcome. For a simple model, I use the following code:

Svy: glm outcome predictor, fam(binomial) link(log) eform

Stata then does some "thinking.....", and takes forever to finally produce a table that has only coefficients and standard errors, but not the risk ratios that I am looking for. Is there something wrong with my data or my code? How would I assess if there' something wrong with the data (eg: convergence, etc?)

I believe I need to use the svy command to incorporate weighting. I was able to obtain risk ratios without the svy command, but I need to use weights. Any suggestions on different commands I could use to get the results I'm looking for?

Thanks in advance,


*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/


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