USE OF PUBLIC RECORDS TO COMPARE RESPONDENTS AND NONRESPONDENTS IN A STUDY OF RECENT WIDOWS

Little is known about bereaved persons who refuse to participate in be reavement research, although participants have been assumed to be repr esentative of the bereaved. This assumption is examined in a study of older women in their first, second, or third year of widowhood. To ide ntify sources of sample bias, data from deceased husbands' death certi ficates (N = 2,331) were used to compare respondents and nonrespondent s at stages in the recruitment process: assumed located, returned repl y card, agreed to hear more, wished to participate, eligible, and inte rviewed. The low participation rate (N = 376) did not produce a major bias, although small, statistically significant differences in decease d husband's age, race, education, place and cause of death were reveal ed. Differences by recruitment stage were observed, as were difference s by cohort. Recent widows were more likely to participate. Implicatio ns for achieving representative samples are discussed.