@joeposaurus It's a bad post. The D-K effect is that people are bad at estimating their own skill, to the point that the regression coefficient between perceived test score and actual test score is very weak (but not zero, or, as in some popular depictions, negative). Mechanically it means people who do worse on tests overestimate their ability and people who do better underestimate it, but that's another way of representing the finding that the regression is meh.