So it’s like, dude!! We got these three groups, right? And we want to see if they’re different from each other. So it’s like, whoa!! What do I do, man? And it’s like, dude! Chill! We can compare the variation between groups to the variation within groups using something called an F-test. Righteous, man! But then you’re like, “Whoa, dude! What if we have measures over time from each subject in each group. Dude, that’s bogus, man!” But that’s not bogus, man! Cuz we can just put in another time to measure variation between times within a subject. So we can measure the differences between groups, the differences between subjects within a group, the differences between times within a subject within a group and the differences within time within a subject within a group. Got all that? Gnarly, dude! And we can assume that the differences between time points within subjects within a group are the same using this thing called compound symmetry or we can that they are different using this bodacious thing called unstructured covariance. This second thing is more real but could also be a bummer if we have too many time points and our model doesn’t want to converge. And that is bogus, man! But no need to be a burnout, bro! Cuz statisticians are coming up with other ways to handle those situations. Righteous! Excellent! So it’s all good, bro!
Okay, after re-reading that, I can see that that was one of the worst explanations of repeated measures ANOVA and one of the worst impersonations of a surfer dude that you ever read in your entire life. Ah well, another one bites the dust. Until next time, hang ten, dudes and dudettes! Okay, I’m stopping now. Promise.