Screening for lung cancer serves to prevent deaths from this disease insofar as earlier resections are associated with higher rates of cure. There is good reason to believe that this is the case: in stage I, the 5-year survival rate with resection is 70%, whereas without resection the corresponding rate is only 10%. Before this evidence emerged, various authoritative organizations and agencies in North America advised against screening for lung cancer on the grounds of the results of several RCTs. As for CXR, I argue that the study results are consistent with up to 40% reduction in the fatality rate. Moreover, modern helical CT screening provides for detecting much smaller tumors than were detected in those studies. It is time to revoke the conclusion that screening for lung cancer does not serve to prevent deaths from this disease, and to quantify the usefulness of CT screening in particular. As for the requisite research, the prevailing orthodoxy has it that RCTs are to be used, but I argue that more meaningful results are obtainable, more rapidly and much less expensively, by the use of noncomparative (and hence unrandomized) studies.