Prev: Macro + Surveyselect
Next: Sorting Problem -
From: SUBSCRIBE SAS-L Anonymous on 24 Oct 2006 20:02
I use OLS to estimate two models of the following form

y=a+bx+bo+bp+bn+?
y=a+bw+bo+bp+bn+?

I want to use Vuong's test to identify whether the difference in the R-
squared between the two models is statistically significant and in short to
identify which model is superior. Sample size: 4,000 firm-year
observations.

I got the predicted values for each model using proc reg but I need some
help on how to compute the log likelihood values.

Further, I want to compare another set of models that have a different
number of independent variables. 6 variables are common to both models, 1
variable is common in both but computed in a different way and one of the
models has 2 additional variables. Any suggestions on how to "correct" the
Vuong test and accommodate for the different number of independent
variables?

Lastly, any other suggestions regarding using a different test that might
be more appropriate in my case are welcome.

Thanks,

Vivian
 | 
Pages: 1
Prev: Macro + Surveyselect
Next: Sorting Problem -