SETTING: Cross-sectional epidemiological study based on a representative sample of the general population in northern Sweden. OBJECTIVES: To assess the prevalence of respiratory symptoms, the role of respiratory symptoms as indicators of impairment of lung function, and to define risk factors for respiratory symptoms and lung function impairment. DESIGN: The 1340 subjects of 6610 who reported respiratory symptoms suggestive of asthma or chronic bronchitis in a postal questionnaire study were invited to a structured interview and lung function tests. A control group of 315 subjects was also invited. Risk factors were assessed from the postal questionnaire. RESULTS: 400 subjects in the symptomatic group had attacks of breathlessness and wheezing, while none in the control group had them, corresponding to 7% of the original study population. Chronic productive cough was present in 537 subjects, of whom 13 were from the control group, suggesting that 12% of the original study population had this symptom. Persistent wheeze was the symptom that predicted the greatest proportion of cases of impaired lung function. Attacks of breathlessness, wheezing, long-standing cough and sputum production were all related to age, smoking and a family history of asthma. Both chronic productive cough and impaired lung function correlated strongly with smoking and age, and their prevalences differed in different socio-economic groups. CONCLUSION: Impaired lung function can be predicted from respiratory symptoms. Data collected in postal questionnaires suffice for the identification of risk factors. Combinations of symptoms gave greater odds ratios than individual symptoms.