[no subject]

[yiwu ye <yiwu21111958@yahoo.com>]

Dear List,

 

I used the following table to show the significant terms in a GLM model in my paper. One reviewer made the following comment:

"Look at the folumn of DF it looks like the model was fitted useing step-wise regression. This distorts the F-statistics since these woudl be F-to-ennter. A more appropriate F would be to start with the full model and calculate F to delete one at a time. This would mean the error residuals are always estimated using the full model".

 

I did this analysis quite some time ago and lost all the codes due to my relacation. My question is when fitting GLMs, can we choose to start with a null model or a full model? Does this really make differences? Thank you so much for your help.

 

Regards,

 

Yiwu

             Df Deviance Resid. Df Resid. Dev P(>|Chi|)
NULL                                31     489.59         
hair             3   165.59        28     324.00 1.138e-35
eye             3   141.27        25     182.73 2.010e-30
sex             1       6.93        24     175.79      0.01
hair:eye       9  146.44        15      29.35 4.806e-27
hair:sex        3     6.27        12      23.08      0.10
eye:sex        3    14.90         9       8.19 1.908e-03
hair:eye:sex  9     8.19         0       1.059e-12      0.52