Stata The Stata listserver
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

Re: st: F test for fixed effect logit?


From   David Harless <dwharles@vcu.edu>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: F test for fixed effect logit?
Date   Wed, 10 Nov 2004 12:57:22 -0500

Robert Bozick wrote:

Hi Stata Users -

I am estimating a fixed effect logit model with persons as my fixed unit. I want to test whether there is sufficient variation among persons to justify the use of a fixed effects model. In a fixed effects model (xtreg), STATA provides an F test following the model testing whether there is variation in the person specific interceptss are equal to zero. I noticed when estimating a fixed effects logit model (xtlogit) that there is no corresponding test.

My two questions are:

(1) Is there a corresponding F test for a fixed effect logit model?

(2) If not, is there a way for me to test whether there is significant variation in the person-specific intercepts in a fixed effect logit framework?

Thanks,
Robert
Robert,
I don't know of a test statistic for xtlogit/clogit analogous to the F-statistic for the significance of the individual intercepts from xtreg. In the fixed effects logit model the "person-specific" intercepts are not estimated at all, they are conditioned out the likelihood function (in the linear regression model xtreg is equivalent to OLS with dummy variables fo the cross-sectional units, but not so with logit).

Greene, Econometric Analysis, 3rd edition, p. 900, addresses this issue and suggests a Hausman specification test (see -hausman-) comparing parameter estimates under the fixed effects logit model to ordinary logit.

Dave Harless
*
* 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/




© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index