Overweighting Private Information: Three Measures, One Bias?