A nice illustrative discussion (here) on various approaches to assess whether "normal" distribution assumptions plausibly hold. Participants in the discussion provide a helpful illustration (below) on two perspectives (kernel density vs. qnorm) on a possible bi-modal distribution.
Comments