reporting model fit in multilevel models for binary response

[sianphillips]

Hi,

I have a simple 2 level model for predicting a binary response, and I am using the second order PQL estimation method. Can anyone help me work out what I should be calculating and reporting in the paper/report about how good the model fit is? I have read all of the MLwiN manuals but can only find reference to R squared being unreliable for binary data.

Many thanks for your time,
Sian

[marloesc]

"Unfortunately because of the way that logit model are estimated in MLwiN through quasi-likelihood, it is not possible to use the usual deviance to compare models. One could use the Intervals and Tests procedures to test individual and sets of estimates for significance. But using MCMC methodology one can compare the overall fit of different models using the DIC diagnostic."

This is in my trainingmanual volume 2.