The stability of confusion: A reply to Mischel and Peake.