A study of the association between spirometry variability and health status in coal miners was conducted. Data obtained in the first round of the National Coal Study, conducted by NIOSH between 1969 and 1971, were used. Data derived from 7790 miners who produced three technically satisfactory spirograms were analyzed. The American Thoracic Society (ATS) criterion and a 200 milliliter (ml) research variability criterion were used to evaluate variability in the spirometric data. According to the ATS criterion, the two best forced vital capacities or 1 second forced expiratory volumes (FEV1s) had to be within 5 percent or 100ml of each other. The research criterion required that the two best FEV1s differ by no more than 200ml. Respiratory morbidity and mortality data of 8475 miners were reviewed. Odds ratios for cough, phlegm, wheeze, shortness of breath, and death were computed for the groups that failed the ATS and 200ml criteria and compared with those who met the criteria. The subjects who failed the ATS criterion had lower mean FEV1s than those who passed and odds ratios for cough, phlegm, wheeze, shortness of breath, and death of 1.75, 1.67, 1.76, 2.71, and 1.30, respectively. The group that failed the 200ml criterion, compared to the group that met the criterion, had a higher mean FEV1 and odds ratios for cough, phlegm, wheeze, shortness of breath, and death of 1.13, 1.07, 1.15, 1.43, and 0.94, respectively. The authors conclude that increased spirometry variability is associated with poorer health, even though the results for the two criteria are different. To minimize the selection bias effects of using the spirometry variability criterion, the current ATS criterion ought to be relaxed to perhaps 10 percent or 100ml.