Questions like this crop up quite frequently,
so search the archives.
A good simple starting-point is to consider that
multicollinearity is a collective property
of the covariates. How the response variable
is measured is at most a secondary issue.
(Translate to locally favored terminology.)
Nick
[email protected]
Richard Williams
> I am not sure why it is "not acceptable" in OLS. How problematic
> multicollinearity is will depend on things like the sample size. If
> multicollinearity is considered problematic for OLS, my guess is that
> it would also be considered problematic for other types of
> regression, such as Poisson.
ALICE DOBSON
>> I am running a simple Poisson model with a set of highly correlated
>> independent variables. It is not acceptable if I were to run an OLS
>> (VIF >>> 10) but is it also not acceptable for a Poisson regression?
*
* 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/
