Digital technology and conventional radiography were compared for use in the interpretation of pneumoconiosis. Conventional radiographs and the International Labor Office standard radiographs, used for classifying pneumoconiosis, were digitized. NIOSH certified B-readers classed the conventional and digitized radiographs over the course of several weeks. Radiograph quality, the profusion, shape and size of small opacities, as well as the presence of pneumoconiosis, were judged. Generally, the digitized images were considered to be of higher quality than the conventional films. The mean small opacity profusion score was similar for the digitized images and the conventional films. Model variability was overwhelmingly due to reader variability or interactions, such as mode versus reader, involving the reader effect. Two of the three readers were significantly more variable than the remaining reader, and also differed significantly in variability from one another. In the determination of pneumoconiosis, two readers exhibited better intrareader agreement using conventional films and the other exhibited better intrareader agreement using the digitized images. Comparing all three readers, better agreement was found for the conventional films. Less variability between display modes was observed in the classifications of small opacity shape and size. Although the digitized images were viewed as being of greater quality than the conventional films, greater variability was found in the interpretation of digitized images. Less variability occurred in later sessions, perhaps indicating a learning effect. The authors conclude that the overall interpretation of pneumoconiosis using digitized images is similar to that using conventional films. However, the digital technology does not appear to offer improvements over the conventional films, such as reduced interreader and intrareader variability.