Thank you, Martin,
I guess that may be the reason, as some of the categories do have
relatively small number of observations. Since the Brant test does not
work, I used the gologit regression, and did Wald tests. Will they
generate the same results as those from the Brant test?
Thank you
Gao
On Sat, Oct 18, 2008 at 9:04 AM, Martin Weiss <[email protected]> wrote:
> as -rnethelp "http://www.indiana.edu/~jslsoc/stata/spost9_ado/brant.hlp"-
> says:
>
> "This is most likely not to be the case with models that have few
> observations in
> the extreme categories and many independent variables."
>
> HTH
> Martin
> _______________________
> ----- Original Message ----- From: "Gao LIU" <[email protected]>
> To: <[email protected]>
> Sent: Saturday, October 18, 2008 2:59 PM
> Subject: st: Brant test
>
>
>> Dear Statalist:
>>
>> I tried to test the parallel slopes assumption of ordered logit
regression
>> using the brant command, and got an error message of "not all independent
>> variables can be retained in all binary logits, brant test cannot be
>> computed". Does anyone know what may be the reason? Thank you very much.
>>
>> Best
>>
>> Gao
>>
>>
>>
>>
>>
>>
>>
>> *
>> * For searches and help try:
>> * http://www.stata.com/help.cgi?search
>> * http://www.stata.com/support/statalist/faq
>> * http://www.ats.ucla.edu/stat/stata/
>>
>
>
> *
> * For searches and help try:
> * http://www.stata.com/help.cgi?search
> * http://www.stata.com/support/statalist/faq
> * http://www.ats.ucla.edu/stat/stata/
>
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
