Rectangular collimation is used in dentistry to reduce radiation by restricting the x-ray beam to approximately the size of a number 2 intraoral film (3.2X4.1 cm). However, this restricted beam size can lead to exposure errors. The aim of this study was to retrospectively evaluate the number of radiographs exposed and the presence of technical errors by the use of traditional circular or rectangular collimators during endodontic therapy on simulated teeth in manikins. A total of 1475 digital radiographs of 84 teeth exposed by 60 dental students were evaluated. Evaluation was done by a board certified endodontist, an endodontic resident, and a dental student. Analysis of the different raters showed no significant differences among the three. Radiographs were randomized and blindly renamed. Repeated-measures mixed-model ANOVA was used to compare the number of radiographs exposed using the different collimators. Although not statistically significant (P<.05), there were 15% more radiographs taken with the rectangular collimator when compared to the circular collimator. Using a repeated-measures logistic regression, there was a significant difference of the proportion of radiographs with cone cuts (P = .0003) taken with a rectangular collimator (59%) compared to radiographs taken with a circular collimator (19%). There was no significant evidence for a collimator difference when considering missed apex (P = .0986) or missed apex due to a cone cut (P = .0631). In order to expose high quality radiographs avoiding cone cuts, a traditional circular collimator may be indicated for use during endodontic therapy.