Hello. I have survey data for 20+ countries and my dependent variable is a dichotomous variable. Scanning around Statalist lead me to the command melogit but since Stata says "initial values not feasible," meqrlogit was presented as an option. Now I am looking for commands on how to test/solve endogeneity problem which I suspect is present. What should I do? Really hoping for a reply. Thank you in advance.