Prev: import variable names from delimited files
Next: Diffplots don't show in GLIMMIX
From: Lindsay on 4 Jul 2007 12:56
Hello,

I just ran a single-equation panel model (N=50, T=180) with TSCSREG. I
received the following note for both the fixone and the fixtwo options
(i.e. one-way and two-way fixed effects models):

"NOTE: The transformed regression does not have full rank. User should be
aware of possible multicollinearity and/or estimability problems before
using FixTwo method results."


As a result, TSCSREG was unable to calculate the Hausman for fixed versus
random effects. I do not, however, receive this note when I run the random
effects model. Is this because of the GLS estimation in the RE model as
opposed to the OLS estimation method in the FE model? Or could my data
layout or program potentially be part of the problem?

In any case, does anyone know how to interpret this NOTE and what its
implications are? Can you suggest any next steps?

Thanks very much,

Lindsay
 | 
Pages: 1
Prev: import variable names from delimited files
Next: Diffplots don't show in GLIMMIX