During the first part of a study that aimed to compare the survival of persons with colorectal cancers, the comparability of data collected by the Geneva Cancer Registry, the Côte d'Or Cancer Registry and the Mallorca Cancer Registry was investigated, as well as the feasibility of obtaining a common reliable stage classification at diagnosis. The validity of incidence and follow-up data was high in the three registries but completeness appeared to be slightly lower in the Mallorca Cancer Registry than in the other two registries. Comparison of incidence curves, by age, showed that a discrepancy appeared over 75 years, which could correspond to an under-diagnosis or an under-registration of some cases among the elderly in the Côte d'Or and the Mallorca Cancer Registries. Stage classification was stratified by surgical treatment in order to improve the homogeneity of investigations undergone by the patients in each class. Stage distribution and stage-specific survival were consistent with those observed in other population-based series. This study shows that it would be better to restrict comparisons of the survival of persons with colorectal cancer to patients under 75 years, and that stage specific survival comparisons are possible with data routinely collected by registries.