Hi,
I have a dataset (T=2 and N=4300 (2300 in the first wave of which 2000 panels are present in the second wave)).
I'm estimating an ordered probit, where the
dependent variable y= takes values between 0 and 2
I wanted to ask you, is it possible to apply the method of first differences if the dependent variable is an ordinal variable and almost all the regressors are dummies?
do I have to consider the difference prime of all the regressors included in the model?
Thank you.

Best,