Re: st: Graphic displays or results from margins

["Weichle, Thomas" <Thomas.Weichle@va.gov>]

Richard,
I'm getting into the margins command as well.  Here is what I've been
able to learn so far.  After producing the predicted probabilities using
the margins command after the logistic regression, I use Roger Newsom's
parmest command to save the results in a dataset and then use the
'eclplot' command to graph the predicted probabilities with the
confidence intervals.  There are some matrices to create after the
parmest command, but I don't find it too cumbersome.  
Try this.  This will produce 4 panels in the figure.  If you feel they
are too cramped together, try using if statements on the command.

webuse nhanes2f, clear
logit diabetes i.black i.female age i.female#c.age, nolog
margins female#black, at(age = (20(10)70)) post

parmest, format(estimate min95 max95) norestore
matrix input age = (20 \ 20 \ 20 \ 20 \ 30 \ 30 \ 30 \ 30 \ 40 \ 40 \ 40
\ 40 \ 50 \ 50 \ 50 \ 50 \ 60 \ 60 \ 60 \ 60 \ 70 \ 70 \ 70 \ 70)
matrix input female = (0 \ 0 \ 1 \ 1 \ 0 \ 0 \ 1 \ 1 \ 0 \ 0 \ 1 \ 1 \ 0
\ 0 \ 1 \ 1 \ 0 \ 0 \ 1 \ 1 \ 0 \ 0 \ 1 \ 1)
matrix input black = (0 \ 1 \ 0 \ 1 \ 0 \ 1 \ 0 \ 1 \ 0 \ 1 \ 0 \ 1 \ 0
\ 1 \ 0 \ 1 \ 0 \ 1 \ 0 \ 1 \ 0 \ 1 \ 0 \ 1)
matrix vars = age, female, black
matrix colnames vars = age female black
svmat vars, names(matcol)

label define female 0 "Male" 1 "Female"
label values varsfemale female
label var varsfemale "Gender"

label define black 0 "Non-Black" 1 "Black"
label values varsblack black
label var varsblack "Race"

eclplot estimate min95 max95 varsage, rplottype(rspike)
estopts(msize(medium)) ciopts(msize(large)) ///
	xtitle("") ///
	by(varsfemale varsblack, style(compact) legend(off)
title("Predicted Probability of Diabetes ({&plusmn}95% C.I.)")) ///
	b2title("Age", color(black) size(medlarge)) ///
	xlabel(20(10)70)  ///
	ytitle("Diabetes Estimated Probability", color(black)
size(medlarge))

eclplot estimate min95 max95 varsage if varsfemale == 0 & varsblack ==
1, rplottype(rspike) estopts(msize(medium)) ciopts(msize(large)) ///
	title("Predicted Probability of Diabetes ({&plusmn}95% C.I.)")
///
	xtitle("") ///
	b2title("Age", color(black) size(medlarge)) ///
	xlabel(20(10)70)  ///
	ytitle("Diabetes Estimated Probability", color(black)
size(medlarge))


Tom Weichle
Math Statistician
Center for Management of Complex Chronic Care (CMC3)
Hines VA Hospital, Bldg 1, C202
708-202-8387 ext. 24261
Thomas.Weichle@va.gov 


*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/