Predicting distress-induced eating with self-reports: mission impossible or a piece of cake?