Hi everyone
I would like to generate lists of groups of variables that have been
generated via xi from the column names of the vector e(b) after
regression commands.
It's probably easier to explain with an example...
sysuse cancer
stset studytime, fail(died)
gen randcat=1
replace randcat=2 if uniform()>1/3
replace randcat=3 if uniform()>2/3
xi:stcox i.drug age i.randcat
. matrix list e(b)
e(b)[1,5]
_Idrug_2 _Idrug_3 age _Irandcat_2 _Irandcat_3
y1 -1.7300519 -3.2313217 .12791436 -.93457115 -.77850524
I would like to take the column names from e(b) to generate a varlist
of the xi-generated variable groups, something like:
_Idrug* _Irand*
...so that I can loop through these and run testparm on each group:
. testparm _Idrug*
( 1) _Idrug_2 = 0
( 2) _Idrug_3 = 0
chi2( 2) = 25.98
Prob > chi2 = 0.0000
. testparm _Irand*
( 1) _Irandcat_2 = 0
( 2) _Irandcat_3 = 0
chi2( 2) = 3.57
Prob > chi2 = 0.1678
I am using stepwise procedures++ with xi-generated variables entered
as a group e,g,
stcox (_Idrug*) age, so I don't know in advance which variables will
be included in a model.
Thankyou in advance
Karin
++I am aware of the problems associated with stepwise regressions, but
I am trying to get an exploratory grip on a large dataset!
*
* 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/
