Temporal trends in disease are often caused by changes in exposure or health care and modeling these trends may be an early step in beginning a study of the etiology of disease. Once some understanding of etiology has been achieved, it is useful to compare quantitatively the agreement between what has been learned from analytical research and the experience in the population. The age–period–cohort framework provides a way to analyze the agreement between an analytical epidemiological study and population rates, summarizing limitations on the age, period, and cohort scales. Lung cancer provides a classical example of a disease that has age, period, and cohort trends and it is primarily caused by cigarette smoking. A carcinogenesis model that allows for trends in smoking in U.S. males is described, as is the age–period–cohort model used for calibration and assessment of model adequacy. Results from this analysis show that a large portion of disease trends are accounted by smoking trends using this model, but there are limitations that might be the result of model limitations or limitations in the exposure data.