Goodness-of-fit test for multilevel log-binomial regression

[Bree Shi]

Hi Statalist Users,

I am just wondering what is the best GOF test for multilevel log-binomial regression? Is H-L appropriate?
My code is as follows

gllamm depvar var1 var2 , i(site ) family(binomial) link(log)  eform

Also attached H-L graph below.
Graph.gph

I have seen suggestions on using Hjort–Hosmer statistics, but not sure how to do it in Stata.

Thank you for you help in advance!

Many Thanks,
Bree

[George Ford]

https://www.stata.com/meeting/australia15/abstracts/materials/oceania15_quinn.pdf

[Bree Shi]

https://www.stata.com/meeting/australia15/abstracts/materials/oceania15_quinn.pdf

Hi George,

Thank you! I have been trying to find the hh.ado program the authors write but to no avail. I am very new to this. Do you have any direction as to where I would be able to assess that? I have tried Stata search.

Many Thanks,
Bree

[Andrew Musau]

If the authors did not make the program public, it will be difficult to get hold of it. Why not just email them?