Abstract Sexual size dimorphism (SSD) has been extensively documented in many organisms, from vascular plants to vertebrates, including mammals. Geometric morphometric methods offer new, powerful tools for the study of SSD. The investigation of shape variability has not been previously pursued in traditional SSD studies and is an unique contribution of geometric morphometrics to this field. The combining of Procrustes methods with multivariate statistical analyses supports all traditional SSD experimental designs and allows new studies of shape variability and its relationship with size and other variables. Intraspecific and interspecific examples comparing traditional and geometric morphometric approaches to the study of SSD within several mammalian taxa are presented.