Jessica Kiefte-de Jong --
The only way to know the "effect" at 6 months is to include data at 0
months or before, so that the variables you have included are not
collinear. Note also that the command you have specified is
essentially just OLS (i.e. equivalent to -regress-) and ignores any
panel structure (that -xtreg- or xtivreg2- might exploit). As for
showing coef estimates for all dummy variables, you may be interested
in -devcon- (from SSC).
webuse nlswork, clear
qui tab year if union<., g(dtime)
reg wks_ue union dtime2-dtime12, robust
devcon, groups(dtime*)
est store devc1
xtgee wks_ue union dtime*, i(idcode) corr(independent) robust nolog
devcon, groups(dtime*)
est store devc2
est table *, se
On 4/20/07, Jessica Kiefte-de Jong <[email protected]> wrote:
Thanks a lot, Maarten. But how can I know if there is an effect of the
intervention at 6 months? For all the other time periods I'll use
interaction terms (e.g. Dtime12*intevention) but I cannot perform that for
the 6 months dummy.
Jessica
--- Jessica Kiefte-de Jong wrote:
> I am using GEE in STATA for analysis of an interventionstudy where
> data collection was performed every 6 months and I am assessing time
> as categorial variable by using dummies. However, STATA dropped the
> dummy I used as reference catagory (Dtime6). How can I fix this problem?
There is no problem to solve. Stata does exactly what it should be doing.
An observation can fall into one of four categories: 6, 12, 18, and 24
months after intervention. The dummies tell you whether or not a case is
in category 6, 12, 18 or 24. If you know that a case is not in categories
12, 18 and 24 (i.e. Dtime12 Dtime 18 Dtime24 are all 0) than we know that
it must be in category 6. If we know the case is in one of the other
categories (i.e. one of Dtime12, Dtime 18, Dtime24 is 1) we know that that
case is not in category 6. As a result, the dummy Dtime6 contains no
information that isn't present in the other dummies and must be removed.
The remaining dummies will measure how different a case is relative to
period 6.
Hope this helps,
Maarten
*
* 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/
