Both models have 2main+4 controls+time dummies. However, in one of the two models I have industry dummies and in the other one I do not ( because adding industry dummies is leading to a singular matrix error). That is the only difference between the two RE models. Do I have to conduct some test to prove that the co-efficients are comparable (because of difference in the number of predictors between the two models (I am using EViews8 for data analysis).....
Also, what if I just avoid the industry dummies in the second model...bcos, if I don't add them to the second model, the results are fine (no error is being produced). Can I do that??? Will it effect the compaprablity of the two models
Well, if one model produces an error, the question of the comparison is not really meaningull: the model with an error is not interpretable, unless you show that this error does not prevent the convergence of the model and is harmless...
- Badges
- Science topic