The aim of this study was to investigate the extent to which differences between physically light and heavy work, with respect to the development of health problems, might in cross-sectional data be eliminated or diminished by health-related selection into and out of occupations. The study population comprised 12,314 men and 12,393 women between 25 and 74 years of age, who were interviewed within the framework of the Statistics Sweden Survey of Living Conditions. For all respondents, detailed occupational histories were recorded. For all the diseases and functional disorders studied, morbidity differences were apparently diminished compared to what would be expected if there was no selection. This association was most apparent when studying musculoskeletal disorders, impaired hearing and impaired working capacity for men.