--- "Friedrich Huebler <huebler@r...> wrote:
> Is it possible to add different text to the individual plots in a
> by() graph? In a response to a question by Joseph Wagner I posted
> these commands two days ago:
>
> . sysuse auto
> . count if mpg<. & weight<. & foreign==1
> . local for = string(r(N), "%2.0f")
> . count if mpg<. & weight<. & foreign==0
> . local dom = string(r(N), "%2.0f")
> . scatter mpg weight, by(foreign) text(40 1000 "N = `for'") text(40
> 4000 "N = `dom'")
Unfortunately this method didn't work for me, I did not get sample sizes:
count if bmi<. & age<. & hivst==1
gen sample = r(N) if hivst==1
count if bmi<. & age<. & hivst==2
replace sample = r(N) if hivst==1
sum age if bmi<. & hivst<.
gen xpos = r(max)
sum bmi if age<. & hivst<.
gen ypos = r(max)
twoway (scatter bmi age) (scatter ypos xpos, mlabel(sample) msymbol(i)),
by(hivst, legend(off))
The method Nick Cox described did give me my sample sizes but I now have
been asked to do the same for box plots that are over twice.
graph box bmi, medtype(line) over(sex, label(labsize(small))) over(hivst)
box(1, bfcolor(none)) ///
box(2, bfcolor(none)) blabel(total, position(inside) format(%9.1f))
-hivst- and -sex- are numeric with value labels
I was wondering about the blabel option. Is it possible to get blabel to
produce the count for that particular box?
*
* 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/
