Hello
I am doing a survey logistic regression from a DHS dataset.
I have tried to use the stata command svylogitgof for goodness of fit and the linktest as the diagnostics.
I have two alternative models - in one the svylogitgof proves to be significant but the linktest shows there is no specification error and in the other model its exactly the opposite. Both tests dont show me favourable results at the same time.
I am running out of time trying to solve this problem!! is there anything that I am missing or that can be done about it?
I am doing a survey logistic regression from a DHS dataset.
I have tried to use the stata command svylogitgof for goodness of fit and the linktest as the diagnostics.
I have two alternative models - in one the svylogitgof proves to be significant but the linktest shows there is no specification error and in the other model its exactly the opposite. Both tests dont show me favourable results at the same time.
I am running out of time trying to solve this problem!! is there anything that I am missing or that can be done about it?
Comment