Goodness of Fit Test(s) for Zero-Inflated Model Using Complex Survey Data

[Dana Dumont]

I'm running a series of zero-inflated Poisson models with complex survey data using the svy command. To test goodness of fit, I've been running the models without the svy command and calculating the AIC and BIC to compare between models, however, I know that with weighted data these stats would probably shift. Are there any goodness of fit tests within Stata that will work for both zero-inflated models and complex survey data concurrently? Any insight would be greatly appreciated!

[Dana Dumont]

Update: I've been trying the chi2gof command (https://www.stata-journal.com/articl...article=st0360) but so far to no avail...

I'm getting the error: "matsize must be between 10 and 800". Using Stata IC 16.1 where the maximum matrix size is 800, my model has 20-25 variables in the inflation portion and 10-15 in the count portion. Same error appears when I remove the svy command. Not sure why this error is occurring...