June 2017
Beginner to intermediate
576 pages
15h 22m
English
When looking at the diagnostic plots, you are only looking at one possible distribution of the residuals. The library TeachingDemos has an interesting interactive function named vis.test which will challenge you to pick out the residuals which were generated from the regression which was run, with an alternate set of randomly generated residuals from the same regression. Remember, the residuals are supposed to be random, so the test essentially randomizes the residuals. If you are able to discern your own residuals from the other residuals that it generates, the residuals fail the test, and the test concludes (with a p-value of course) that your residuals really aren't random, which ...