...In truth, we need a little bit of both, but that is just an opinion of a statisician. Econometrics have several major divisions:
one of which is generalized linear regression models. We need to worry about variance/co-variance structure of the regression residuals in order for us to make i) unb... (read more)