A user wrote privately (edited to Statalist conventions)
we are using the command -ciplot- from SSC and we have a question: is it
possible
to plot also the geometric mean with its confidence interval?
Otherwise do you have suggestion to solve our problem?
-ciplot- does not support geometric means. But doing it yourself is an
option.
sysuse auto, clear
gen ub = .
gen lb = .
gen mean = .
levelsof rep78, local(levels)
qui foreach l of local levels {
means mpg if rep78 == `l'
replace ub = r(ub_g) if rep78 == `l'
replace lb = r(lb_g) if rep78 == `l'
replace mean = r(mean_g) if rep78 == `l'
}
scatter mean rep78, pstyle(p1) || ///
rcap ub lb rep78, legend(off) ytitle(`: var label mpg') pstyle(p1)
*
* 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/
