Random effects are error effects in a layered design (like split plot) that cause correlation between observations, and introduce complications in the expectation of the mean squares. JMP finds the expectations in terms of variance components, solves to estimate the variance components, and constructs denominators for F-tests that have the right expectation.
Analyze: Fit Model.
Highlight and select “random effect” in the Effect Attribute popup menu.
