STUDY OBJECTIVE--The aim was to evaluate the appropriateness of postal questionnaires in studies of spontaneous abortion in a general population. DESIGN--Survey of spontaneous abortions by comparing diagnoses reported in questionnaires to information collected from hospital records and a discharge register. SETTING--The study took place in the general community. PARTICIPANTS--The study involved a random sample of 1400 women born between 1935 and 1960, living in two towns in the south west of Sweden. MEASUREMENTS AND MAIN RESULTS--9% of spontaneous abortions among respondents were not reported in the questionnaire, and 31% were missing in the register; 80% of the spontaneous abortions reported in the questionnaire could be verified in hospital records, and a further 7% were mentioned in the records but not certified by a doctor; 97% of the diagnoses received by the women who had answered affirmatively to two complementary questions on spontaneous abortions could be found in the records. CONCLUSIONS--Postal questionnaires can be used in epidemiological studies of spontaneous abortions, among a general population as well as among occupational groups.