Hi Statalisters,
I am running `teffects ipwra` for causal inference, but the regression does not converge. I found that if I remove a categorical variable, it converges.
Upon inspection, this categorical variable has 30 levels, many of which have fewer than 10 observations.
I must not combine some of the levels of this categorical variable. Moreover, I can't exclude if from the regression.
Can anyone suggest how to proceed? Propensity score matching is essential for my analysis.
Thanks.
