Most likely it means that some characteristics do not vary over
countries, and thus can be perfectly predicted with country dummies.
Note that introducing dummies does not produce fixed effect regression
in non-linear case like -oprobit-.
On 9/17/07, Chris O'Keefe <[email protected]> wrote:
> Hi all-
> I'm running ordered probit models on project level foreign aid data.
> I'm running multiply imputed datasets through clarify to get my
> predicted probabilities. I've run my base models, and I want to check
> them against a model with country dummies. When I attempt this, I get
> a "matrix not positive definite" error message. I've done this with
> using robust and with standard errors clustered by country. Here's my
> code:
>
> estsimp oprobit enviroimpact2 wbcofinancer wbcofyr ldevcom
> lwbenvirocount lepiisdb lngo ligdp ligdpgrowth liexportsgnp lischlsec
> lipotwat lisulfdiox lithrtndmamm licoallandarea liorgpollperh20
> liforest lpop lpolity2 lcorrupt lgovstab _Jcount*, robust mi(isdbno)
>
> where _Jcount* is shorthand for my country dummies, created previously with xi.
>
> Is the solution similar to this* thread? I.e., do I simply need to
> ensure that my ado files are properly updated?
>
> I'd appreciate any help on this.
>
> Thanks,
> Chris O'Keefe
>
> *http://www.stata.com/statalist/archive/2006-10/msg00097.html
> *
> * 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/
>
--
Stas Kolenikov, also found at http://stas.kolenikov.name
Small print: Please do not reply to my Gmail address as I don't check
it regularly.
*
* 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/
