The short answer to your first question is because this is the nature of the
FE logit model.
Indeed it only uses observation that switch from one period to the other (1
to 0) or (0 to 1).
A comprehensive survey of fixed effects panel data models is in Honor�
(2002). Or you can look at
Chamberlain, G. (1985): "Heterogeneity, Omitted Variable Bias, and Duration
Dependence,"
in Longitudinal Analysis of Labor Market Data, ed. by J. J. Heckman, and B.
Singer, no. 10
41in Econometric Society Monographs series,, pp. 3-38. Cambridge University
Press, Cambridge,
New York and Sydney.
anna
-----Original Message-----
From: [email protected]
[mailto:[email protected]] On Behalf Of Levy Lee
Sent: 27 October 2004 10:28
To: [email protected]
Subject: st: fixed effect logit model
Dear Statalister,
Just a quick and maybe naive question, why the sample size shrinked when i
estimate a fixed effect logit model using an unbalanced panel dataset?
the question goes like this:
i firstly estimate estimate an ordinary pulled data logit model with a
three year panel of 3107 observations for 1241 firms. when i tried the fixed
model, the observation number dropped to 1045 observations, which only
captured 427 firms. can you tell me why?
also, why these is no a similar xtprobit model for fixed effect?
thanks
Levy
*
* 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/
*
* 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/
