The use of combined mood scores in studying the premenstrual syndrome: does it clarify or muddy the picture?