What’s the difference between fixed and random effects?


Gelman (2005, p. 21) to the rescue.

We prefer to sidestep the overloaded terms “fixed” and “random” with a cleaner distinction […]. We define effects (or coefficients) in a multilevel model as constant if they are identical for all groups in a population and varying if they are allowed to differ from group to group.

Gelman A. (2005). Analysis of variance—why it is more important than ever. Annals of Statistics, 33(1), 1–53


One thought on “What’s the difference between fixed and random effects?”

Comments are closed.