Lung cancer mortality caused by exposure to residential radon (10043922) was estimated via the low exposure data collected from underground miners. Data were pooled from 11 cohort studies of underground miners, with a focus on exposures under 100 working level months (WLM). Analyses unrestricted by exposure level were also conducted. Exposed person/years totaled 453,604 in the less than 50WLM group, 564,772 in the less than 100WLM group, and 888,906 in the unrestricted data group. WLM means were 19.7WLM in the less than 50WLM group, 40.0WLM in the less than 100WLM group, and 493.6WLM in the unrestricted data group. The relative risks (RRs) increased with increasing cumulative exposure, from 1.00 at 0WLM to 1.86 at 86.6 to 99.9WLMs. For the linear, excess relative risk (ERR) model, ERRs of 0.0117 per WLM for the less than 50WLM group, 0.0080 per WLM for the less than 100WLM group, and 0.0044 per WLM for the unrestricted data group were estimated. The linear ERR model generally fit the observed RRs well. Time since exposure, age, and exposure rate and duration significantly affected the unrestricted data, but not the restricted data. The model incorporating time since exposure, age, and exposure duration fit the less than 100WLM data better than did the model incorporating time since exposure, age and working level (TSE/AGE/WL). Both models fit the less than 50WLM data equally well. The models were applied to residential radon concentration data. Lung cancer risk for males ranged from 0.078 based on the linear ERR model with the less than 100WLM data to 0.141 based on the TSE/AGE/WL model with the unrestricted data. The authors conclude that for estimating lung cancer mortality induced by residential radon exposure via exposure to miners, the results of restricted data analyses are similar to those of unrestricted data analyses.