From: "Nick Cox" <[email protected]>
Reply-To: [email protected]
To: <[email protected]>
Subject: RE: st: ordered probit with random effects
Date: Tue, 14 Feb 2006 10:45:38 -0000
As often pointed out on this list, "doesn't [...] work"
conveys no precise information.
As I understand it, -reoprob- is not on SSC,
so -ssc- cannot install it. Nor is it
an official Stata command, so Stata
will not recognise the command unless
it is installed.
You must install the -reoprob- files from the STB-61
files on the Stata website. Type
findit reoprob
and within the window that will open
click on
sg158.1
and then click to install.
Nick
[email protected]
Roys Nicolas
> Thanks a lot for your reply.
> in fact, the problem is even more basic.
> Stata9 doesn't recognize the command : reoprob
> so i try to install it (ssc install reoprob) but that doesn't
> work either.
Joseph Coveney
> 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/
