The methacholine challenge test protocol was assessed in the reanalysis of three occupational studies. We evaluated the impact of truncating the range of methacholine on responsiveness, as defined by slope and PC(20). In original analysis, reactivity was similar for apprentices and auto body shop workers, whereas boilermakers were more responsive. Truncating high concentrations did not change the classification of subjects with PC(20) <8 or 16 in any population. However, when responsiveness was measured by slope, the mean responsiveness increased, from -7.9 to -15.3 for apprentices and -7.2 to -10.0 for auto-body shop workers. Results support the American Thoracic Society's recommended maximum of 16 mg/ml and provide evidence that extending the dose range beyond that does not increase sensitivity, whereas stopping before 16 may exaggerate response. Furthermore, to ensure validity, neither slope nor PC(20) should be extrapolated beyond data.