Objective –Several studies have compared attenders and non-attenders in organised cervical screening programmes but few have analysed subgroups of at-tenders and non-attenders. This study presents social and other characteristics of such subgroups. Setting—Aarhus County, Denmark. Methods –A case-control study in a cohort of 133 500 women, aged 23–60, included in the programme from 1 October 1990 to 1 April 1994. The participation rate was 75%, and those taking part comprised women with opportunistic screening or who had had a smear owing to symptoms in the previous three years (“active” attenders), and women who were invited for screening because they had not been otherwise tested (“passive” attenders). “Passive” (n = 708) and “active” attenders (n = 692) were compared. Women who had never had a smear test (“never” attenders, n = 287) were then compared with “ever” attenders (n = 1215)—that is, women who had not had a smear test during the previous 42 months, but had had at least one previous test. Data were collected by mailed questionnaires. Results –The response rate was 81% and 53% for attenders and non-attenders, respectively. After correction for age, there was no difference between the “active” and “passive” attenders for cancer risk factors (smoking, age of first intercourse, number of sexual partners, and social group), or in the degree of responsibility for close relatives, but “active” attenders seemed to have more frequent contact with their general practitioner. “Never” attenders had less frequent contact with their general practitioner than “ever” attenders. They were more often living alone and nullipara, but had no overrepresentation of cancer risk factors. Conclusions –Increased effect cannot be obtained by focusing on the described groups, but by increasing the participation rate. “Never” attenders do not belong to a special risk group.