[Limdep Nlogit List] Random Effects - Results and Het Models
Christopher McIntosh
cmcintos at d.umn.edu
Fri Jul 20 02:37:55 EST 2007
I am running a pure random effects model (using random in command to
suppress fixed effects) and have two questions:
1. What are the results in the "Least Squares with Group Dummy
Variables" dialog box telling me? Since the fixed effects model is not
being run are these results for the random effects model?
2. When I add a heteroscedasticity specification (either
;heteroscedasticity or ;het = group) the bottom of the "Least Squares
with Group Dummy Variables" dialog box has this message "White
estimator was not P.D. Using OLS matrix", what is P.D. and what does
this statement imply about my results?
See attachment for an example.
Thank you in advance for your help,
Chris
Christopher McIntosh
Assistant Professor
Department of Economics
SBE 165, Room 171
Duluth, MN, 55812
218-726-8408
cmcintos at d.umn.edu
-------------- next part --------------
+----------------------------------------------------+
| Least Squares with Group Dummy Variables |
| Ordinary least squares regression |
| Model was estimated Jul 19, 2007 at 10:04:53AM |
| LHS=WTP Mean = 100.6068 |
| Standard deviation = 476.7052 |
| WTS=none Number of observs. = 5738 |
| Model size Parameters = 1994 |
| Degrees of freedom = 3744 |
| Residuals Sum of squares = .5438834E+09 |
| Standard error of e = 381.1404 |
| Fit R-squared = .5828222 |
| Adjusted R-squared = .3607507 |
| Model test F[***, 3744] (prob) = 2.62 (.0000) |
| Diagnostic Log likelihood = -41018.83 |
| Restricted(b=0) = -43527.03 |
| Chi-sq [***] (prob) =5016.40 (.0000) |
| Info criter. LogAmemiya Prd. Crt. = 12.18459 |
| Akaike Info. Criter. = 12.15440 |
| Estd. Autocorrelation of e(i,t) -.244134 |
| White estimator was not P.D. Using OLS matrix. |
+----------------------------------------------------+
More information about the Limdep
mailing list