Introduction Self-ratings of health (SRH) are the most widely used single-indicator of health in many scientific disciplines as well as in epidemiology. Accordingly, many researchers rely on SRH in describing, examining, and explaining (social) inequalities in health. How these ratings come about, however, is still relatively unexplored. For instance, little is known about the extent to which undiagnosed diseases or the severity of present diseases play a role in rating one's health. Biomarkers, such as HbA1c for diabetes, play an interesting role in this regard since they give objective and highly relevant health information beyond what a respondent is willing - or able - to tell. Thus, they offer the possibility of investigating the relevance of diseases unknown to respondents or the severity of diseases. Methods Through the example of diabetes and the pertinent lab-measured HbA1c-concentration in the respondents’ blood, this paper analyzes the relative roles of diagnoses and biomarkers in the evaluation of general health. To do so, this paper draws on diverse health data from 2890 respondents aged 50–79 collected in the 2007 and 2009 waves of the Canadian Health Measures Survey (CHMS). The paper quantifies and compares the contribution of self-reported diagnoses of diabetes and biomarker-indicated diabetes, undiagnosed diabetes, prediabetes, and diabetes-severity to explaining SRH. Their contributions are compared between genders, age, and education. Additionally, separate analyses of all 359 self-reportedly diagnosed diabetics in the sample were conducted in order to investigate the role diabetes-severity plays within diabetics when rating their health. Results Across all subgroups, the influence of undiagnosed (pre)diabetes as well as HbA1c on SRH appears to be subordinate to a manifest diagnosis of diabetes. With the exception of older men (aged 65–79), all models show a greater influence of diagnoses than lab-indicated diabetes while lab-indicated prediabetes generally does not affect SRH when controlling for other health indicators. General and subgroup analyses of the influence of severity within diabetics reveal that HbA1c indeed influences diabtetics’ SRH, although only for older respondents (65 + ) as well as respondents with at least post secondary education. Conclusions While further research is needed, these results highlight the importance of diagnoses of chronic diseases in rating one's health. This paper also suggests that SRH is primarily based on information known to and relevant for the daily lives of respondents while undiagnosed diseases and disease-severity are - at least in the case of diabetes - less relevant.