Recent epidemiological studies of the association between lung cancer and exposure to radon and its decay products are reviewed. Particular emphasis is given to pooled case-control studies of residential exposures, and to cohorts of underground miners exposed to relatively low levels of radon. The residential and miner epidemiological studies provide consistent estimates of the risk of lung cancer, with significant associations observed at average annual concentrations of approximately 200 Bq/m³ and cumulative occupational levels of approximately 50 working level months (WLM), respectively. Based on recent results from combined analyses of epidemiological studies of miners, a lifetime excess absolute risk of 5 × 10⁻⁴ per WLM [14 × 10⁻⁵ per (mJh/m³)] should now be used as the nominal probability coefficient for radon- and radon-progeny-induced lung cancer, replacing the previous Publication 65 (ICRP, 1993) value of 2.8 × 10⁻⁴ per WLM [8 × 10⁻⁵ per (mJh/m³)]. Current knowledge of radon-associated risks for organs other than the lungs does not justify the selection of a detriment coefficient different from the fatality coefficient for radon-induced lung cancer. Publication 65 (ICRP, 2003) recommended that doses from radon and its progeny should be calculated using a dose conversion convention based on epidemiological data. It is now concluded that radon and its progeny should be treated in the same way as other radionuclides within the ICRP system of protection; that is, doses from radon and its progeny should be calculated using ICRP biokinetic and dosimetric models. ICRP will provide dose coefficients per unit exposure to radon and its progeny for different reference conditions of domestic and occupational exposure, with specified equilibrium factors and aerosol characteristics.