It may be beneficial to do regression before ANOVA mainly because I think it's easier to explain residuals with simple linear regression (observed-expected). I found it awkward and ran out of time to explain the F-test requirements about normal residuals. It might be easier after regression to understand if they've already seen that tested under regression.
It may be beneficial to do regression before ANOVA mainly because I think it's easier to explain residuals with simple linear regression (observed-expected). I found it awkward and ran out of time to explain the F-test requirements about normal residuals. It might be easier after regression to understand if they've already seen that tested under regression.