Since the human scapula exhibits sexually dimorphic features, it has been examined in detail in recent years, especially in terms of direct anthropometric measurements. During these investigations, only very few authors have used derived variable (area of the glenoid cavity) and they reported that this measurement could be used for estimating sex. However, other areal variables of the scapula have not been studied to determine whether they contribute to predicting sex. Therefore, we aimed to clarify the question of whether derived variables, as well as direct measurements, could contribute to sex estimation. In the present study, four direct scapular measurements were taken from the magnetic resonance images (MRI) and we then calculated two derived variables from 99 male and 105 female adults’ scapulae. Direct measurements are infraspinous fossa height (IFH), maximum scapular breadth (MB), glenoid cavity length (GCL), axillar bone length (ABL), and maximum spine length (MSL). Derived areal variables are MB*IFH and MSL*IFH. Univariate discriminant function analysis yielded overall sex prediction success rates ranging from 82.4% to 67.5%. As a single derived variable, MB*IFH, was the most discriminative for sexing the individuals. Multivariate discriminant analysis yielded 90.5% success in correct sex assign by using three parameters (ABL, GCL, and MSL). These results demonstrate that the analysis of scapular size provides a highly accurate tool for assigning sex.