A new questionnaire (ATS-DLD) developed by the American Thoracic Society (ATS) for the Division of Lung Diseases (DLD) of the National Heart, Lung and Blood Institute, designed to elicit more details about various symptoms and factors associated with chronic obstructive pulmonary diseases, was compared with 2 previous questionnaires, those developed by the British Medical Research Council (MRC) and by the National Heart and Lung Institute (NHLI). From a private census list of residents of Washington County, Maryland, USA, 6 similar population samples of white men and women 25-64 yr of age were selected, one each for administration of the 3 questionnaires by telephone and by mail. A total of 1004 interviews were obtained. Indicated prevalences of chronic cough and chronic phlegm were essentially the same for all 3 questionnaires; the ATS-DLD elicited significantly more symptoms of any wheeze, whereas the NHLI elicited more mild breathlessness. The MRC and ATS-DLD were virtually identical in indicated prevalences of chronic wheeze, asthma and hay fever, conditions not covered in the NHLI. All 3 questionnaires showed similar increases in symptoms with increased cigarette smoking and with decreased forced expiratory volume in 1 s. Among the interviews conducted by telephone, the NHLI showed a greater variation in symptomatology by interviewer than either the ATS-DLD or MRC, although sampling variation could not be ruled out as a contributory factor. Among the questionnaires self-administered by mail, the ATS-DLD showed the lowest percentage of misunderstood questions, with NHLI next and MRC significantly higher.