it can be difficult to test hypotheses about dissociative identity disorder, because it is an especially rare disorder. which research design would be most useful in these circumstances?