The variety of characteristics related to odontoma research, including an unexplored one such as size, merits a multivariate approach that allows the adequate drawing of inferences with pertinent conclusions. The objective of this study is to establish the possible association between some characteristics related to the odontoma, tumor size among them. The sociodemographic characteristics of 60 patients were evaluated. Diagnosis, size, location, type of treatment performed, and prognosis were determined. These data were analyzed descriptively and through multivariate models. Thirty-four compound and 26 complex odontomas in 32 men and 28 women were observed. The age average of patients was 15.6±11 years. Most of the odontomas presented a size inferior to 10 mm. A statistically significant association was observed between the routine radiographic finding and the absence of dental eruption (p=0.0001). The model of linear regression adjusted between odontoma size and age (?=0.321, p=0.01), as well as the model of logistic regression adjusted between gender (men) and tumor size (OR=12; 1.7 - 93 IC 95%, (p=0.01) were statistically significant. Statistically significant associations between odontoma size and age, and between the male gender and odontomas smaller than 10 mm were found adjusting by other confounding variables. These results could grant clinicians a greater knowledge of the context of odontoma characteristics, which in turn could favor a better diagnostic and therapeutic decision-making.