Epidemiologic investigations of orthodontic treatment should include both objective clinical measures and elements of perceptual assessment in the definition of malocclusion and skeletal disproportion. The effect of dental training and the view of the face presented as a stimulus on judgments of facial attractiveness were evaluated using a method recommended by Howells & Shaw (1) for epidemiologic surveys. Three views (two full face and one profile) of 18 orthodontic patients were presented as stimuli to three panels of judges with different levels of dental training (16 orthodontic residents, 17 dental students, and 71 undergraduate students). Ratings for facial attractiveness were obtained using a visual analog scale. The visual analog scores given the series of 54 slides by each judge were then ranked to create a "location-free" outcome measure. Both the ratings and the rankings of these ratings differed significantly among the three views for 80% of the patients. However, no view was consistently rated or ranked as most attractive across all patients. The ordering of the views from least to most attractive for a given patient appears to be highly dependent on the patient being presented as stimulus. Future studies should consider showing multiple views of a subject simultaneously if the intent is to obtain an overall treatment need score that incorporates assessment of facial attractiveness. The patients were consistently rated as more attractive by the orthodontic residents than by either dental or undergraduate students. When the ranks were analyzed, there were significant differences between the panels but the consistent pattern of differences between the panels disappeared.(ABSTRACT TRUNCATED AT 250 WORDS)