Katsuhide Isa
>
> Recently I wrote the following simple program
> titled 'coefvar.ado' for the purpose of calculating
> coefficients of variation.
>
> ----------coefvar.ado----------
> program define coefvar, rclass byable(recall)
> version 7
> qui summ `1', det
> local coefvar = r(mean)/r(sd)
> di in green _col(8) `coefvar'
> end
> ----------coefvar.ado----------
>
> But somehow it doesn't work as expected.
> More specifically, when typing like
>
> 'bys year : coefvar basepay'
>
> using a data set containing multiple panels, following
> results:are shown on the screen:
>
> __________________________________________________________________
> __________
> ___
> -> year = 1989
> 3.7137332
>
> __________________________________________________________________
> __________
> ___
> -> year = 1990
> 3.7137332
>
> __________________________________________________________________
> __________
> ___
> -> year = 1991
> 3.7137332
>
> __________________________________________________________________
> __________
> ___
> -> year = 1992
> 3.7137332
>
> __________________________________________________________________
> __________
> ___
> -> year = 1993
> ....
>
> and so on. Namely, they are all the same for
> each year when they should vary from year
> to year.
> I guess something is wrong or missing with this
> program, but I'm sorry I can't fix it.
Gary Longton replied to the question directly.
Two ther comments are that
sd / mean
is the definition of coefficient of variation in the literature
I have encountered and that
egen mean = mean(basepay), by(year)
egen sd = sd(basepay), by(year)
gen cv = sd / mean
tabdisp year, c(cv)
gets you there interactively.
Nick
[email protected]
*
* 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/
