Dear all,
I am using the instrumental variable approach. My X variable is a binary variable, and Y is an ordinal variable. However, OLS suggests that Y does not meet the parallel test, so I would better take Y as a categorical variable or try to convert Y into binary variables. How to convert the ordinal variable (e.g., bad, moderate, good) into binary variables in order to perform IV regression? Or does anyone know how to perform IV regression with an ordinary variable but does not meet the parallel test?
Thanks for your help.
0 Response to Instrumental variable with ordinal dependent variable(reject the parallel test)
Post a Comment