Two sets of longitudinal pulmonary function data have been analyzed to determine the effects of parental smoking on the development of lung function in children and adolescents. One community population was obtained in East Boston, Massachusetts, and one in Tucson, Arizona; both studies started in the 1970s and are still continuing. These data sets, analyzed by different methods, have yielded different answers in regard to the effects of parental smoking. A common analytical approach is now used to determine whether the different outcomes were due to the different analytical methods previously used. These results were compared with results from a similar parallel analysis conducted by the East Boston investigators and indicate that the differences found were not due to the methods used for analysis. Reasons for the differences found are explored. The most likely factor responsible for the disparate results is the exposure difference in the two populations, related to proportion and amount of maternal smoking and different indoor environments.