Following the Chernobyl accident, radioactive fission products, including (131)I and (137)Cs, were deposited in Bryansk Oblast in Russia. Intakes of radioiodines, mainly (131)I in milk, were the principal sources of radiation doses to thyroids of residents of the contaminated areas, but those radionuclides decayed before detailed contamination surveys could be performed. As a result, (137)Cs deposition density is the primary measure of the contamination due to the accident and there are relatively few measurements of the ratio of (131)I to (137)Cs in vegetation or soil samples from this area. Although many measurements of radiation emitted from the necks of residents were performed and used to estimate thyroidal (131)I activities and thyroid doses, such data are not available for all subjects. The semi-empirical model was selected to provide a dose calculation method to be applied uniformly to cases and controls in the study. The model was developed using dose estimates from direct measurements of (131)I in adult thyroids, and relates settlement average thyroid doses to (137)Cs contamination levels and ratios of (131)I to (137)Cs. This model is useful for areas where thyroid monitoring was not performed and can be used to estimate doses to exposed individuals. For application to children in this study, adjustment factors are used to address differences in age-dependent intake rates and thyroid dosimetry. Other individual dietary factors and sources (private/public) of milk consumed are reflected in the dose estimates. Countermeasures that reduced thyroid dose, such as cessation of milk consumption and intake of stable iodine, are also considered for each subject. The necessary personal information of subjects was obtained by interview, most frequently of their mothers, using a questionnaire developed for the study. Uncertainties in thyroid dose, estimated using Monte Carlo techniques, are presented for reference conditions. Thyroid dose estimates for individual children made using the semi-empirical model and questionnaire data compare reasonably well with dose estimates made for 19 children whose thyroid burdens of (131)I were measured from May to June 1986.