In a cross-sectional study of 2363 schoolchildren living in two rural areas of New South Wales, we used a questionnaire to collect details of sex, area of residence, social class, early respiratory illness (ERI), parental history of asthma and recent upper respiratory tract infection (URTI), and we used skin-prick tests to measure atopic status. The relative importance of these factors on the likelihood of children having bronchial hyperresponsiveness (BHR) was assessed using a linear modelling analysis. The extent to which these factors affected the severity of BHR was also examined. We found that social class or recent URTI had no association with BHR, that sex and area of residence (inland or coastal) had a small association and that a history of early respiratory illness, a history of asthma in either parent, and atopic status had an important association with BHR. Atopic status was the most important factor. The proportion of children with atopy, with ERI or with parental asthma increased as the severity of BHR increased. The odds ratio for moderate or severe BHR doubled if either ERI or parental asthma was present in addition to atopy and there was a six-fold increase if all three factors were present together. The identification of these risk factors makes it possible to predict which children in the community are most likely to have BHR, and which children are at high risk for having more severe levels of BHR.