I thing you did. I have one comment and one question.
Comment: A nice trick to label the years for which you have data is
. levelsof year, local(lab)
. twoway rbar lower upper wageclass, xlabel(`lab')
Question: Adjusting the bar-width with
. levelsof wageclass, local(K)
. local width = (`:word 2 of `K'' - `:word 1 of `K'')
. twoway
|| rbar lower upper wageclass, barwidth(`width') horizontal
does not work sufficiently for me. On my screen (Stata 10.1 for Linux), the bars overlap,
(although I think they shouldn't). In my proposal I have therefore used
. local width = (`:word 2 of `K'' - `:word 1 of `K'') * .95
to scale the bar width somewhat further down. Wasn't that necessary under Windows?
Uli
Am Samstag, den 21.03.2009, 17:18 +0100 schrieb Antoine Terracol:
> Thanks to Ulrich, Nick and all the others who answered my question, I
> think I met the challenge...
>
>
> Antoine
>
>
> For anyone interrested, here is the code:
>
> /*----------------------------*/
>
> clear
> set scheme s2mono
> webuse nlswork.dta, clear
> g wage=exp(ln_wage)
> su wage, de
> drop if wage>r(p99)
> gen t=(year)
> g ti=t
> g ti2=t^2
> g ti3=t^3
> qreg wage ti ti2 ti3
> est store med
> qreg wage ti ti2 ti3 , quantile(.95)
> est store p95
> qreg wage ti ti2 ti3 , quantile(.05)
> est store p5
>
> drop ti ti2 ti3
> g ti=t+runiform()
> g ti2=ti^2
> g ti3=ti^3
>
> est restore med
> predict med
> predict resmed, resid
>
> est restore p95
> predict q95
> predict resq95, resid
>
> est restore p5
> predict q5
> predict resq5, resid
>
>
> by t, sort: egen pct25 = pctile(resmed), p(25)
> by t, sort: egen pct50 = pctile(resmed), p(50)
> by t, sort: egen pct75 = pctile(resmed), p(75)
> egen ufence = adju(resmed), by(t)
> egen lfence = adjl(resmed), by(t)
>
> reg resq5 ti ti2 ti3
> predict resq5hat
>
> reg resq95 ti ti2 ti3
> predict resq95hat
>
>
> sum wage, meanonly
> gen wageclass = autocode(wage,32,r(min),r(max))
> by t wageclass, sort: gen wageperc = _N/r(N)
>
> // rescale wageperc
> sum wageperc, meanonly
> replace wageperc = wageperc/r(max) * .90
>
> gen lower = t
> gen upper = t + wageperc
>
> preserve
> bysort t wageclass : keep if runiform()<0.05 | _n==1
>
> levelsof wageclass, local(K)
> local width = (`:word 2 of `K'' - `:word 1 of `K'')
> twoway ///
> || rbar lower upper wageclass, graphregion(color(white)) ///
> fcolor(gs15) barwidth(`width') horizontal xlabel(68(5)88) ///
> || line med q95 q5 ti, sort lpattern(solid..) lcolor(black..) ///
> legend(off) xtitle(year) saving(b, replace) nodraw
>
>
> twoway ///
> rspike ufence pct75 t , color(black) ///
> || rspike lfence pct25 t , color(black) ///
> || rcap ufence ufence t , color(black) ///
> || rcap lfence lfence t , color(black) ///
> || rbar pct25 pct50 t, barwidth(.5) fcolor(gs15) lcolor(black) ///
> || rbar pct50 pct75 t, barwidth(.5) fcolor(gs15) lcolor(black) ///
> || line resq5hat resq95hat ti , sort lpattern(solid..) ///
> lcolor(black..) graphregion(color(white)) ///
> legend(off) xlabel(68(5)88 , nolabel notick) ///
> fysize(20) saving(b2, replace) nodraw
>
>
> graph combine b2.gph b.gph, graphregion(color(white)) col(1)
> restore
>
> /*----------------------------*/
>
> *
> * 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/
--
[email protected]
030 25491-361
*
* 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/
