Order Change for Stat Methods II: Regression First

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.  
