Perhaps I am not understanding this correctly, but I get an error that I don't expect after adjust.
. xi:xtgee zoster1 ra age65 sex,i(pat) robust eform fam(poisson) nolog exposure(exposure)
GEE population-averaged model Number of obs = 92494
Group variable: patkey Number of groups = 20238
Link: log Obs per group: min = 1
Family: Poisson avg = 4.6
Correlation: exchangeable max = 11
Wald chi2(3) = 36.92
Scale parameter: 1 Prob > chi2 = 0.0000
(standard errors adjusted for clustering on patkey)
------------------------------------------------------------------------------
| Semi-robust
zoster1now | IRR Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
ra | 1.22771 .0940327 2.68 0.007 1.056576 1.426563
age65 | 1.368359 .0760525 5.64 0.000 1.22713 1.525841
sex | .9154671 .0617301 -1.31 0.190 .8021324 1.044815
exposure | (exposure)
------------------------------------------------------------------------------
. adjust age65 sex,by(ra) exp ci
offset (ln(exposure)) not constant within by()
However,
. tab exposure,m
exposure | Freq. Percent Cum.
------------+-----------------------------------
.5 | 92,635 100.00 100.00
------------+-----------------------------------
Total | 92,635 100.00
shows that the exposure is constant.
Am I misunderstanding something?
Thanks,
Fred Wolfe
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
