The rate of childhood thyroid cancer incidence observed in northern Ukraine during the period 1986-1998 is described as a function of time-since-exposure, age-at-exposure, and sex. Conclusions are drawn for the excess absolute risk per dose: after a minimal latency period of about three years it shows a linear increase with time-since-exposure for at least nine years. It is roughly constant in age-at-exposure, up to 15 years. For girls exposed very young it is about a factor 2 larger than for boys. For children exposed at age 16-18 this ratio increases to about 5. The thyroids of young children are not more sensitive to radiation dose than those of older ones in absolute risk in northern Ukraine in the currently used data set. As the background is increasing with age, a constant absolute risk gives a decreasing relative risk.