Because the SAS output shows the SCALE parameter, I think that Ariel
reported the model-based, not the empirical, standard errors. If I
recall, model-based standard errors ignore the clustering (REPEATED
statement) and so, are just OLS. The ratio of the total sample size
to the number of clusters is about 25:1, with a square root of 5:1.
This would account for some, but not all, of the standard-error
discrepancies. Ariel, please show the second part of the GENMOD
results.
-Steve
On Wed, Jul 22, 2009 at 1:30 PM, Joseph Coveney<[email protected]> wrote:
> Ariel Linden wrote (excerpted):
>
> Here is the code I used for the comparative models GLM in Stata vs GEE in SAS.
>
> [output redacted]
>
> --------------------------------------------------------------------------------
>
> No answers--just a few observations:
>
> 1. You've got Stata and SAS fitting different models (compare the values for
> scale). SAS is fitting with an exchangeable working correlation, and Stata is
> fitting with an independent working correlation. Is the exchangeable working
> correlation essential? If not, then try TYPE=IND or CORR=IND on the REPEATED
> statement in PROC GENMOD.
>
> 2. PROC GENMOD prints out two sets of coefficients and standard errors, as I
> recall, one with "model based" and the other with "empirical" standard errors.
> The latter is what I assume you're showing. (The table's header got cut off in
> the cut-and-paste).
>
> 3. Use [pweight = iptw], not [aweight = iptw]. It couldn't hurt to use the
> correct weighting scheme.
>
> Joseph Coveney
>
>
>
> *
> * For searches and help try:
> * http://www.stata.com/help.cgi?search
> * http://www.stata.com/support/statalist/faq
> * http://www.ats.ucla.edu/stat/stata/
>
--
Steven Samuels
[email protected]
18 Cantine's Island
Saugerties NY 12477
USA
845-246-0774
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
