Proc Logistic - AIC, SC and parameter estimates

[mate_25]

Hi.

I've been doing some logistic regressions using Proc logistic.
When I compare the estimates, P-values and the Model Fit Statistics to those i get, from doing the same logistic regression in Eview, i get very different results. My results from SAS however, match the ones i get from SPSS.

For instance:
SAS: AIC=601, SC=605
Eviews: AIC=1.097, SC=1.14 and

Also, a lot the variables that are significant in Eviews are not the ones that are significant in SAS. I find this quite strange.


Does anyone have an idea of what goes wrong or what might cause the differences?

Thanks in advance.

Code in SAS:
proc logistic data=mroz;
model inlf (event='1') = nwifeinc educ exper age kidslt6 kidsge6;
run;

[Mean Joe]

I don't use Eview, so I'm not much help. Other than trying to make some guesses.

Why are the significant variables different? In SAS, it only uses records where all the variables are non-missing (inlf nwifeinc educ exper age kidslt6 kidsge6). Do you have records with maybe one of these variables missing but all others filled in? Might want to check then if Eview does something to include missing values.