Ionizing radiation is a well-known risk factor for cancer. Unresolved issues still exist in the quantitative assessment of cancer risk in the low and medium dose range which is especially relevant for radiation protection purposes. Risk assessments for non-cancer health effects like cardiovascular disease or cataract, and the consequences of exposure to non-ionizing radiation are also not fully understood. We strive to fill knowledge gaps through new epidemiological studies and statistical methods. We conduct studies in populations with diagnostic or therapeutic medical exposure as well as in populations with occupational radiation exposure. Important methodological considerations are the dose reconstruction, identification of confounders, uncertainties in exposure and outcome measurement, and dealing with potential selection effects.