[en] The Hanford data set has attracted attention primarily because of analyses conducted by Mancuso, Stewart, and Kneale (MSK). These investigators claim that the Hanford data provide evidence that our current estimates of cancer mortality resulting from radiation exposure are too low, and advocate replacing estimates based on populations exposed at relatively high doses (such as the Japanese atom bomb survivors) with estimates based on the Hanford data. In this paper, it is shown that the only evidence of association of radiation exposure and mortality provided by the Hanford data is a small excess of multiple myeloma, and that this data set is not adequate for reliable risk estimation. It is demonstrated that confidence limits for risk estimates are very wide, and that the data are not adequate to differentiate among models. The more recent MSK analyses, which claim to provide adequate models and risk estimates, are critiqued. 18 references, 1 table