A range of competing models
The data currently available on the effects of radiation is mostly taken from studies on survivors of Hiroshima and Nagasaki, and is riddled with large errors due to lack of statistics. The graph on the left shows how, for doses between 0 and 3 sieverts, the rate of the 'linear relationship with no threshold' can be determined. The margins of errors do not allow us to distinguish between this straight line, a straight line with a threshold effect and an intermediary model. The magnified view on the right shows the range of doses under 500 mSv, where the risks predicted by the three models vary significantly. In this domain of low doses, it seems not possible to decide between the models, even if their predictions differ considerably..
IN2P3