Statement of problem. Quantitative measures of occlusal contacts are of paramount importance in the study of chewing dysfunction. A tool is needed to identify and quantify occlusal parameters without occlusal interference caused by the technique of analysis. Purpose. This laboratory simulation study compared occlusal contacts constructed from 3-dimensional images of dental casts and interocclusal records with contacts found by use of conventional methods. Material and methods. Dental casts of 10 completely dentate adults were mounted in a semi-adjustable Denar articulator. Maximum intercuspal contacts were marked on the casts using red film. Intercuspal records made with an experimental vinyl polysiloxane impression material recorded maximum intercuspation. Three-dimensional virtual models of the casts and interocclusal records were made using custom software and an optical scanner. Contacts were calculated between virtual casts aligned manually (CM), aligned with interocclusal records scanned seated on the mandibular casts (C1) or scanned independently (C2), and directly from virtual interocclusal records (IR). Sensitivity and specificity calculations used the marked contacts as the standard. Contact parameters were compared between method pairs. Statistical comparisons used analysis of variance and the Tukey-Kramer post hoc test (P=<.05). Results. Sensitivities (range 0.76-0.89) did not differ significantly among the 4 methods (P=.14); however, specificities (range 0.89-0.98) were significantly lower for IR (P=.0001). Contact parameters of methods CM, C1, and C2 differed significantly from those of method IR (P<.02). The ranking based on method pair comparisons was C2/C1 > CM/C1 = CM/C2 > C2/IR > CM/IR > C1/IR, where ">" means "closer than." Conclusion. Within the limits of this study, occlusal contacts calculated from aligned virtual casts accurately reproduce articulator contacts.