Studies of observer performance with non-manipulated digital images are almost lacking; nor has the effect of cathode ray tube (CRT) display formats been evaluated. The present study compared the diagnostic accuracy of observers with and without clinical experience in evaluating intra-oral dental films and the corresponding non-manipulated digitized images. Three different-sized (5", 9", 17") monitors were used to study the effect of CRT display format. Artificial radiolucent and radiopaque lesions in dry human mandibles were examined. ROC curve areas were averaged and cross-tabulated according to viewing method and observer experience. For the total group, the non-manipulated digital images resulted in smaller ROC curve areas than the plain films for all CRT formats, with a significantly inferior result for the 17" images (p = 0.036). The increase in diagnostic accuracy of the digital images with decreasing CRT size was not significant (p > 0.05). There was some difference in the mean ROC curve areas between teachers and students, but the effect of the viewing methods was not significant. In terms of observer performance, the experienced observers demonstrated a reduced, and the inexperienced an increased, variability with the digital images compared with conventional film.