I'd like to compare the fit of 2 models which predict vote choice in election.
Supposed that samples (voter) have 5 alternatives (candidates), and they feel
different utility for an each alternative.
Models have 4 variables ( age, gender, education, adn utility), but utility is
calculated in a different way across 2 models. Samples are same.
I did clogit and got some measures of fit as follows:
Model A
Measures of Fit for clogit of choice
Log-Lik Intercept Only: -1046.135 Log-Lik Full Model: -735.931
D(634): 1471.863 LR(16): 620.407
Prob > LR: 0.000
McFadden's R2: 0.297 McFadden's Adj R2: 0.281
Maximum Likelihood R2: 0.615 Cragg & Uhler's R2: 0.615
Count R2: 0.538
AIC: 2.314 AIC*n: 1503.863
BIC: -2634.538 BIC': -516.775
Model B
Measures of Fit for clogit of choice
Log-Lik Intercept Only: -1046.135 Log-Lik Full Model: -740.903
D(634): 1481.806 LR(16): 610.463
Prob > LR: 0.000
McFadden's R2: 0.292 McFadden's Adj R2: 0.276
Maximum Likelihood R2: 0.609 Cragg & Uhler's R2: 0.609
Count R2: 0.557
AIC: 2.329 AIC*n: 1513.806
BIC: -2624.594 BIC': -506.832
Questions:
// Which model of them does a good prediction?
// How do I get the adjusted count Rsquare in clogit?
// How do I get a cross table of the observed X the predicted values ?
( 'lstat' has already been installed, but does not work well...)
// What is the best measures of fit in clogit?
Naoko Taniguchi
Dep. of P.S., University of Michigan
----
TANIGUCHI, Naoko [email protected]
*
* 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/
