> It's hard for many to accept that their ability to read other people is only as accurate as flipping a coin, but that's what we've got evidence for.

Does anyone have any links that support this? Thanks if so.