Logit Regression Interpretation

#1
All,

Im new to this site, so thanks for having so much info here!!

I have just done analysis using logit regression and was wondering if anyone could help with the interpretation of the results, specifically:

1. If one of the goodness of fit tests fails (p<0.05), is the model invalid?
2. If the concordant percentage is less than 75%, can the model still be used?

Any help or any websites that anyone knows which help with this question would be great.

Again, thanks and I look forward to more posts!!

Paul
 
#3
Thanks for the reply.

No, its not always the Hosmer and Lemeshow, I have one that has failed the Pearson and one that has failed the Deviance.