Abstract: BACKGROUND: People with a lower socioeconomic position have a higher the prevalence of most self-rated health problems. In this article we ask whether this may be attributed to self-rated health not reflecting actual health, understood as mortality, in different socioeconomic groups. MATERIAL AND METHOD: For the study we used data from the Nord-Trondelag Health Study 1984-86 (HUNT1), in which the county's entire adult population aged 20 years and above were invited to participate. The association between self-rated health and mortality in different occupational classes and income groups was analysed. The analysis corrected for age, chronic disease, functional impairment and lifestyle factors. RESULTS: The association between self-rated health and mortality was of the same order of magnitude for the occupational classes and income groups, but persons without work/income and with poor self-rated health stood out. Compared with persons in the highest socioeconomic class, unemployed men had a hazard ratio for death that was three times higher in the follow-up period. For women with no income, the ratio was twice as high. INTERPRETATION Self-rated health and mortality largely conform to the different socioeconomic strata. This supports the perception that socioeconomic differences in health are a reality and represent a significant challenge nationally. Our results also increase the credibility of findings from other studies that use self-reported health in surveys to measure differences and identify the mechanisms that create them.