Cook's Distance for Multilevel Mixed-Effects

[Charles Ehmat]

Hey everyone,

I am using Stata 16 on Mac.

I am running a 2-level mixed model (until now with only random intercepts but I will add random slopes next). My final model is this:

mixed recovery c_trust c_ltrust senior_org ib0.b03evac ib0.intens c.c_age##c.senior_org c.c_brgyears##c.prep_measures || villageID:, reml residuals(independent, by(intens))

Now I want to run influence diagnostics, especially Cook's Distance and DFBETAS to find outliers and influential observations. Is there a way to run the

predict newvarname, cooksd

command after using the mixed command also?

Any tips are greatly appreciated!