First, note that -esta- and -esto- have been renamed. The new names
are -esttab- and -eststo-.
Second, note that the exponentiated coefficients can be tabulated
using the -eform- option. This also transforms the confidence
intervals.
Now the answer to the question: -esttab- will not be able to collect
the labels for the xi-dummies from the original variable. Therefore,
you'd have to label the xi-variables and use the -label- option in
-esttab-. Furthermore, the reference category can be included using
the -refcat()- option (documented in -help estout-). Here's an
example:
. sysuse cancer, clear
(Patient Survival in Drug Trial)
. quietly stset studytime, failure(died)
. eststo: xi: quietly stcox i.drug
i.drug _Idrug_1-3 (naturally coded; _Idrug_1 omitted)
(est1 stored)
. lab var _Idrug_2 "Tadalafil"
. lab var _Idrug_3 "Sildenafil"
. esttab, eform wide ci nostar label refcat(_Idrug_2 "Placebo")
----------------------------------------------
(1)
_t
----------------------------------------------
Placebo ref.
Tadalafil 0.231 [0.0894,0.596]
Sildenafil 0.0707 [0.0213,0.235]
----------------------------------------------
Observations 48
----------------------------------------------
Exponentiated coefficients; 95% confidence intervals in brackets
. esttab, eform wide ci nostar label refcat(_Idrug_2 "Placebo", label(1))
----------------------------------------------
(1)
_t
----------------------------------------------
Placebo 1
Tadalafil 0.231 [0.0894,0.596]
Sildenafil 0.0707 [0.0213,0.235]
----------------------------------------------
Observations 48
----------------------------------------------
Exponentiated coefficients; 95% confidence intervals in brackets
. eststo clear
ben
On 6/11/07, Jannik Helweg-larsen <[email protected]> wrote:
( Stata 9.2+ estout- fully updated:)
Just discovered Benn Janns extremely useful commands esto/esta.One
question for esta:
Is it somehow possible to make esta keep/inherit the variable/value
labels after using xi e.g in line with what " reformat" (by Tony
Brady) does?
Example (needs estout and reformat installed from ssc):
sysuse cancer
label define drug 1 "Placebo" 2 "Tadalafil" 3 "Sildenafil"
label val drug drug
stset studytime, failure(died)
esto: xi: stcox i.drug
estadd expb
esta, main(expb) wide ci nostar
--------------------------------------
(1)
_t
--------------------------------------
_Idrug_2 0.231 [-2.415,-0.518]
_Idrug_3 0.0707 [-3.849,-1.449]
--------------------------------------
N 48
--------------------------------------
expb coefficients; 95% confidence intervals in brackets
I would prefer to get the following type of output, in which the
original labels for drug are inherited:
xi: stcox i.drug
reformat
-----------------------------------------------------------------------------
Covariate Haz. Ratio Std. Err. P>|z| 95% Conf. Interval
-----------------------------------------------------------------------------
Drug type (1=placebo)
Placebo* 1
Tadalafil 0.231 0.112 0.002 (0.089 to 0.596)
Sildenafil 0.071 0.043 <0.001 (0.021 to 0.235)
-----------------------------------------------------------------------------
* Baseline category
bw
Jannik
*
* 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/
*
* 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/
