2

I am new to statistics and I am looking for a general answer. Can someone explain (very simply) what is the relationship between Fixed-effect & Random-effect models with Generalized Linear Models?

Are Generalized Linear Models a type of fixed-effect model? I am having a hard time differentiating and understanding these three concepts.

Steffen Moritz
  • 1,564
  • 2
  • 15
  • 22
Ber Baysan
  • 21
  • 1

1 Answers1

1

The proposed duplicate addresses what fixed and random effects are well.

Generalized linear models are unrelated to these terms, though both fixed & random effects can be used in them. They are easiest to understand with reference to linear models, especially regression. They differ primarily in allowing for residual distributions that are more appropriate that the normal (e.g. for binary data, or discrete data), and also for the variance to depend on the mean.

mkt
  • 11,770
  • 9
  • 51
  • 125
  • 1
    Thank you! Just have one more question. What is the difference between Pooled OLS and FE Models? I will be using panel data, and I know that a Hausman test will determine if I should used a FE model or RE model. But how do I know if I should use pooled OLS or FE model, for example? – Ber Baysan Jul 22 '19 at 08:47
  • 1
    @BerBaysan I suggest posting that as a new question with more context. – mkt Jul 22 '19 at 08:48