Model fitness indices for logistic regression with svyset

[Philip Baiden]

I'm using data from the BRFSS and running logistic regression models in blocks (4 blocks). My challenge is that when I declare the data as survey data using svyset, I do not get model fitness indices such as pseudo R-square and -2 log-likelihood to enable me to assess if the set of predictors included in subsequent models are contributing to model improvements. When I run the model without svyset, I get those model fitness indices. Is it proper to run the models with svyset and report the results, but for model fitness purposes, run the models without svyset and just obtain the fitness indices?

[George Ford]

https://www.stata.com/manuals/svyestat.pdf

estat gof should work.