The study was aimed to assess: a) the proportion of true refusers among non-attenders resulting from residents and screening database matching, b) the typology of refusers compared to attenders, and c) the reasons for non-attendance as stated by refusers. Three hundred and forty-one non-attenders according to computer-produced lists were interviewed: 29 were not evaluable, 148 had had a Pap test, 18 had not attended for valid reasons, and 146 were true refusers; 94 refusers accepted the interview as far a points b) and c) were concerned. Ninety-four attenders, matched by age and residence, were interviewed for comparison. Attendance at private laboratories or incorrect identification of subjects on the screening record accounted for 45% or 55% of misclassified cases, respectively. The adjusted true refusal rate in the last 10 years was 18.1% compared to a computer-assessed non-attendance rate of 40.7%. Refusal was significantly associated to socioeconomic status (geographic origin, working activity and educational level of both women and husbands), whereas no significant association was observed for sexual history (deliveries, age at first sexual relation, number of partners) or health-related behavior. Attendance and counseling at the gynecologist's office were strongly associated to attendance, and gynecologists should be stimulated to promote opportunistic screening. The belief that screening is useless in the absence of symptoms, the fear of cancer being detected, postponement or laziness were the reasons for non-attendance stated by 80% of refusers. Improving knowledge about the purpose of the Pap test should be the main goal of any action aimed to improve screening attendance.