I am using -simulate- to run a program I wrote in
order to estimate statistical power. I have several
outcomes, and for most of those outcome variables
there is no problem.
However, for three of the variables, I find that after
one iteration, -simulate- hangs. (I let it hang for two
days on three occasions, so I think it's not just
my impatience.)
The program is fairly simple, it runs -xtmixed- and for
a while I thought it was -xtmixed- that was hanging; however,
a few well placed -display- commands disabused me of that
notion. In the below I get two dots (".") displayed before
it hangs:
=======================================================
program define xtpow, rclass
version 10
syntax, out(string) eff(real) [, indvars(string) ]
u power, clear
bys id_practice id_doctor: egen mean=mean(`out')
expand 2
bys obs: gen cohort=_n-1
replace mean=mean+`eff'
replace mean=min(mean,1)
replace mean=max(mean,0)
rndbinx mean denom
replace `out'=bnlx if cohort==1&!missing(`out')
di in r ". " _con
xtmixed `out' cohort `indvars' || id_practice: || id_doctor:
matrix b=e(b)
matrix V=e(V)
scalar zrat=b[1,1]/sqrt(V[1,1])
scalar pval=2*(1-normprob(abs(zrat)))
di in r ". " _con
return scalar p=pval
end
=======================================================
My command:
. local O "chol_ldl_good"
. local E=0.005
. simulate pval=r(p) , reps(1000) : xtpow, out("`O'") eff(`E')
produces
chol_ldl_good
. .
Like I said, this works when the local macro O is 12 out of
15 variables, but not for the other three. There is nothing
strange about those three - and the simulation always completes
ONE iteration with no trouble, it just never starts the second.
Thoughts?
Jeph
*
* 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/
