The stability of individual differences in gender typing: Implications for understanding gender segregation