You are possibly missing out on an
easier solution from Roger Newson's
-eclplot-.
Nick
n.j.cox@durham.ac.uk
Fredrik Wallenberg
> My advisor wanted me to plot the coefficients for some category
> variables with "whiskers" for the standard errors. The program below
> does the trick but I'm still left with one issue... I haven't figured
> out how to get the variable names on the x-axis rather than a number.
> The names are stored in the e(b) matrix as row names, but I haven't
> figured out how to get them into a variable.
>
> An example of the output can be found at:
>
> http://sims.berkeley.edu/~fredrik/files/IncReg.jpg
>
>
> #delimit;
> ** plot estimation coefficients ** ;
> capture program drop graphbeta ;
>
> program define graphbeta ;
> version 8.2 ;
> args varname ;
> quietly { ;
> matrix beta = e(b)' ;
> matrix var = vecdiag(e(V))' ;
> local varnames : rownames(beta) ;
> svmat float beta, names(beta) ;
> svmat float var, names(betavar) ;
> gen lowbound = beta1 - 2*sqrt(betavar1) ;
> gen hibound = beta1 + 2*sqrt(betavar1) ;
> egen varcount = fill(1 2) ;
> };
> twoway (rspike lowbound hibound varcount if beta1!=.,
> blwidth(medthick) blpattern(solid) blcolor(maroon)) (scatter beta1
> varcount if beta1!=., mcolor(navy)), ytitle(Coefficient +/- 2 s.e.)
> yline(0) xtitle(`varname' Category Coefficients) legend(off)
> graphregion(fcolor(white) lcolor(white) lwidth(none)) ;
> quietly {;
> matrix drop _all ;
> drop beta1 - varcount ;
> };
> end ;
> ** ;
*
* 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/