Background: Many patients with resectable non-small cell lung cancer (NSCLC) are unfit for lobectomy owing to comorbidity. Surgical outcomes are biased by preoperative selection factors and upstaging that occurs during surgery. This study compares outcomes between sublobar pulmonary resection and traditional external beam radiation therapy. Methods: This cohort study utilizes Surveillance, Epidemiology, and End Results-Medicare data (1998 to 2005). Patients with stage IA NSCLC treated with either radiotherapy or sublobar resection were compared. The bias of clinical staging in the radiation group versus pathologic staging in the surgical group was addressed by including only sublobar resections without lymph node sampling. Medicare claims data were used to calculate a modified Charlson comorbidity score for each patient. Results: In all, 878 patients received radiotherapy and 657 underwent sublobar resection without lymph node sampling. Radiation patients were older (77.0 versus 75.5 years, p < 0.0001) and had larger tumors (22.8 versus 17.9 mm, p < 0.0001). There was no difference in comorbidity scores between groups (p = 0.21). Three-year overall survival favored sublobar resection (56% versus 35%; p < 0.0001). Predictors of earlier death were radiation, age, comorbidity score, tumor size, male sex, and prior malignancy (all p < 0.05). Propensity analysis matched 319 radiation patients and 319 sublobar resection patients. In this subgroup, 3-year overall survival favored sublobar resection (52% versus 41%; p < 0.001). Conclusions: Sublobar resection without lymph node sampling appears to be superior to radiotherapy for clinical stage IA NSCLC. For patients with prohibitive risk for lobectomy, sublobar resection may be preferable to radiotherapy. Radiotherapy results in current and future patients are likely to be better.