I like to ask this whenever Dunning Kruger comes up; DK was a sample of Cornell undergrads, and the study was one tenth the size of the Stanford study. DK participants were volunteers of a psych class who got extra credit. Presumably they needed and could use extra credit, which may have excluded the A and the F students. It’s hard to imagine ways to start with more bias, or how we can possibly accept this sample as representative of humanity.