What are the assumptions to be tested and reported when we use Fixed-effect and Random-effects models?  The same assumptions such as normality,  no autocorrelation, and no heteroscedasticity?

Thank you.

Lema