Some methodological problems with retrospective studies of the relationship between childhood bereavement and later behavior disorders are reviewed and some results of a followup study of a large sample of 9th grade, Minnesota students are presented. (WL)