The various criticisms of our test of the linear no-threshold theory of radiation carcinogenesis in the paper by Smith et al. are considered and shown to be invalid. It is shown that there is no significant difference between the BEIR IV formula and the formula we use, that the uncertainties in effective average radon exposures in U.S. counties due to the issues they raise are not very large and that even if they were implausibly large, the results of our study would not be much affected. I review the seven essentially independent methods we used to estimate smoking prevalence, all of which give the same results but most of which, including the most important, were ignored by Smith et al.; explaining our results by uncertainties in smoking data would require correlations between radon and smoking that are grossly implausible. Our use of measurements of radon, smoking, and lung cancer rates from different time periods is justified, and it is shown that if more recent lung cancer rates are used, the results are not changed. Problems in comparing Iowa data with our study are discussed. It is shown that many of their criticisms of our study are more applicable to the case-control and cohort studies that they endorse. Many of their conclusions are presented without valid supporting evidence. A simple procedure is suggested that can easily settle any questions about the validity of our study; with this procedure, I offer to show that any other published ecological study might give invalid results. The point here is that our study is very different from all other published ecological studies.
(C)1998Health Physics Society