Thanks to Roger and Kaleb. Kaleb's solution works, but only if you know the
valuelabelname. Ordinarily that will not be known, so I didn't take that path.
Or Fred could have written the malfunctioning line as
local `catvar'lab`num': label (`catvar') `num'
(note the parentheses), which refers to the label of the variable stored
in the macro -catvar-. See -[P] macro- for details.
Thanks! I missed the parenthesis when I read the text. I saw the
parenthesis as part of the formatting of the - [P] macro page rather than
as part of the necessary code. Thanks, the program works now.
Or, alternatively, Fred might like to use the -factext- package,
downloadable from my website. (A fully up to date version should be on SSC
as soon as Kit gets back to Boston.) The -factext- package inputs the
-label- variable in a -parmby- output and re-creates the categorical
factors, sometimes with help from the -descsave- package, which is also on
SSC. The categorical factors can then be used in plots and tables.
I am indebted to Roger for his parmest program, but took a different tack
in displaying the results for several reasons. First, I use -tlabels-, a
secondary labelling system (see -fsum- ssc install fsum,replace). Roger's
program are not designed for that. The program I am working on uses tlabels
if they are present, otherwise uses regular labels. Second, I wanted a
simple program for non-sophisticated users for whom the utilities Roger
provides might seem too complicated. Third, I wanted some special
formatting of confidence intervals that are not available in Roger's superb
series of programs. Here is an example of running my new program (yet
unfinished).
Thanks again to Roger.
. parmby "xi:logistic married haq pain sex fatigue age edlevel ", label
eform format(estimate min95 max95 %8.2f p %8.1e) saving(killme,replace)
Command: xi:logistic married haq pain sex fatigue age edlevel
Logistic regression Number of obs = 7415
LR chi2(6) = 349.84
Prob > chi2 = 0.0000
Log likelihood = -4300.4191 Pseudo R2 = 0.0391
------------------------------------------------------------------------------
married | Odds Ratio Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
haq_disabi~y | .8428295 .0411647 -3.50 0.000 .7658894 .9274987
pain_scale | .9860523 .0132516 -1.05 0.296 .9604188 1.01237
sex | 2.829347 .2200756 13.37 0.000 2.429277 3.295303
fatigue_sc~e | 1.003245 .0121197 0.27 0.789 .9797697 1.027283
age | .9805909 .0021429 -8.97 0.000 .9763998 .9848
edlevel | .999307 .011803 -0.06 0.953 .9764393 1.02271
------------------------------------------------------------------------------
file killme.dta saved
. showparmest using killme, sort ciform(to)
Variable Odds Ratio 95% C.I. P value
Sex (% male) 2.83 (2.43 to 3.30) 8.9e-41
Age (years) 0.98 (0.98 to 0.98) 3.0e-19
HAQ (0-3) 0.84 (0.77 to 0.93) 4.6e-04
Pain (0-10) 0.99 (0.96 to 1.01) 3.0e-01
Fatigue (0-10) 1.00 (0.98 to 1.03) 7.9e-01
Education (years) 1.00 (0.98 to 1.02) 9.5e-01
Constant 8.50 (5.37 to 13.46) 7.0e-20
file killme.dta saved
I hope this helps.
Best wishes
Roger
Fred Wolfe
National Data Bank for Rheumatic Diseases
Wichita, Kansas
Tel (316) 263-2125 Fax (316) 263-0761
[email protected]
*
* 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/
