SAS Proc Logistic Selection=Score -

i'm using proc logistic procedure in sas , option selection=score gives me few logistic regression models , chi-square values. question model better - smaller chi-square or bigger?

in general, larger chi-squared statistic correspond lower p-value (more significance). however, important know shape of chi-squared distribution , number of degrees of freedom. can see in graph, relationship between p , chi-squared changes based on degrees of freedom.