Roys Nicolas wrote:
I want to estimate an ordered probit with random effects. I use Stata9 but
the command reoprob doesn't seem to work. Could smbdy give me the right
procedure to use ? (i find little guidance in previous post nor in google).
--------------------------------------------------------------------------------
-reoprob- is suitable for random effects ordered probit regression. Could
you be a little more specific about your problem?
Try the little do-file below. It converges by the fourth iteration (null
model and full model each--about half a minute to one minute in total) and
should for you, too. If it does, then something is might wrong with the way
you set up your analysis (or you're not being patient enough for your
dataset). If it doesn't, then something might be wrong elsewhere: is
everything up-to-date (official ado files, -reoprob-, executable)? It
should give you some diagnostic information either way.
Joseph Coveney
clear
set more off
drawnorm latent0 latent1 latent2, ///
corr(1 0.5 0.5 \ 0.5 1 0.5 \ 0.5 0.5 1) ///
n(200) seed(`=date("2006-02-14", "ymd")')
generate float predictor = uniform()
generate int row = _n
reshape long latent, i(row) j(time)
generate byte manifest = 1
forvalues cut = 0.2(0.2)0.8 {
replace manifest = manifest + (norm(latent) > `cut')
}
xi: reoprob manifest i.time*predictor, i(row)
exit
*
* 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/
