The aim of this in vitro study was to compare cone-beam computed tomography (CBCT) to conventional radiography (RG) in the assessment of the periodontal ligament space. A phantom with a variable "artificial" periodontal ligament space (0, 100, 200, 300, and 400 mu m) was used as a model. The examinations were performed simultaneously with RG and NewTom (R) 9000 digital volume tomograph. Assorted after increasing widths, 15 RGs and 15 CBCT images were presented for judgment to 20 dentists (DD), 20 dental assistants, and 20 dental students. Several weeks later, the same images were randomly mixed and presented to the same 20 DD again. The trial shows that RG gaps wider than 200 mu m could be correctly identified by all participants with an accuracy of nearly 100%. A significant difference was observed between the modalities (p < 0.05 and p < 0.001) where conventional RGs performed better than CBCT for assessment of periodontal ligament space. Interobserver variation in relation to each technique was evaluated and no significant difference was found (p > 0.05). In subjective evaluations of image quality with CBCT, the results were basically inferior for images of artificial periodontal ligament space, regardless of the experience of the observers.