The following is from a research paper that I recently read:
The model includes country fixed-effects to allow any trends in voting behavior to vary from country to country (without fixed-effects, the results are unchanged).
My questions are:
Doesn't any model allow trends in voting behavior to vary? Voting behavior is the dependent variable in the model. So of course it's going to change since the time series of the data is 30 years.
What does it mean that without fixed effects, the results are unchanged?
