Background: Self-reports are frequently used to assess occupational exposures in epidemiological studies on asthma, but the validity and influence of asthma status on performance is unclear. Methods: Data on self-reported exposure to air pollutants were obtained for 16,752 randomly selected working individuals from 40 study centers, and compared to exposures obtained by a job exposure matrix (JEM). The influence of current asthma symptoms or medication was investigated. Results: Specificity of self-reports amounted to 0.83 and 0.87, and sensitivity 0.48 and 0.42 for asthmatics and non-asthmatics, respectively, when compared with the JEM. Self-reported exposure, but not exposure assessed by the JEM, was more prevalent in areas with a higher community prevalence of asthma. Conclusions: The prevalence of self-reported occupational exposures seems to depend on asthmatic health status at both the individual and the community level. Associations between self-reports and asthma are likely to be biased, especially in pooled analyses combining different areas with varying prevalence rates of asthma. © 2005 Wiley-Liss, Inc.