The reproducibility of answers to childhood respiratory symptom questions was investigated by administering two childhood respiratory symptom questionnaires twice, with a one month interval, to the same population of Dutch school children. The questionnaires were completed by the parents of 410 children from 6-12 yrs old. One questionnaire was developed by the World Health Organization (WHO). The other questionnaire was developed for use by school health services and was based mainly on the American Thoracic Society (ATS) questionnaire for respiratory symptoms in childhood. It was called the Child Health Care (CHC) Questionnaire. Half of the parents received a questionnaire in which the WHO questions came first on both occasions, the other half a questionnaire in which the CHC questionnaire came first on both occasions. The reproducibility of the answers was assessed by calculating Cohen's Kappa, a chance-corrected measure of association. Questions about "wheeze", "asthma", "bronchitis", "cough and phlegm", and "pneumonia" were found to give good to excellent reproducibility, with Kappa values ranging from 0.60-0.96. Questions about "phlegm", "cough", "chest congestion" and "runny/congested nose" were less satisfactory in this respect, with Kappa values generally well below 0.60. The reproducibility of the answers did not depend on the order in which the questionnaires were administered, on the gender or age of the child, or on the smoking habits or educational level of the parents.