In the R Doc we do comparison of deviance of 2 models... then in the SAS Doc we do comparison of 2 models using backward elimiation. THey give slightly different P-values.. (maybe because SAS is using - 2 * Log LIklihood instead of deviance? Needs checking & writing up.
Needs to edit the below & add inot the comparison page if any difference in found.
NOTE: the chi-square test summary of backward elimination, p=0.6250 is different to the results in R, which gave a difference in deviance of -0.24046, p=0.6239. This difference is currently being investigated.
In the R Doc we do comparison of deviance of 2 models... then in the SAS Doc we do comparison of 2 models using backward elimiation. THey give slightly different P-values.. (maybe because SAS is using - 2 * Log LIklihood instead of deviance? Needs checking & writing up.
Needs to edit the below & add inot the comparison page if any difference in found.
NOTE: the chi-square test summary of backward elimination, p=0.6250 is different to the results in R, which gave a difference in deviance of -0.24046, p=0.6239. This difference is currently being investigated.