Link: A demonstration of an issue with ETTR
This link requires a bit of explanation (if only so that I can remember it later). The person I'm linking to took a standard colour checker test target and took a properly exposed shot then a series at increasing positive exposure compensations (ie, exposing to the right), postprocessed all of the overexposed photos to correctly expose them again, and then cropped strips of the colour targets and stacked the same strips from each exposure. The goal is to clearly see any colour shifts caused by ETTR.
(As he notes, some are deliberately overexposed, going beyond what you should theoretically do with ETTR. Of course they do represent what happens if you accidentally overexpose in the course of trying to do ETTR.)
Part of what I like about this is that it's an experiment that anyone can do (if you have a colour checker test target). Your camera and processing system may well give you different results than his, but either way you'll have learned something interesting.
PS: obviously you should do this test with the camera on a tripod and locked at a single ISO, unless you also want to test the effects of ISO on colour shifts. Although that too may be a useful test, depending on how you're thinking of using ETTR.
Written on 31 October 2011.