In The Blank Slate, Pinker talks about research on stereotyping and says it was found that stereotypes do tend to be correct statements about average properties of groups. That is, they represent "mean field" theories, in physics terms. Of course, that means they don't apply to every member of the group, so in that sense they're inaccurate, because they miss the variation within the group i.e. that not every member of the group exhibits the average properties of the group.