<p>With greater availability of post-mortem computed tomography (PMCT), the Brooks and Suchey method is frequently applied in medico-legal contexts using 3D reconstructions of the pubic symphysis. The scientific literature contains reference samples constructed from either dry bones (autopsy series) or 3D reconstructions (virtual samples). This study aims to evaluate the impact of reference sample nature (dry bone vs. virtual bone) on age estimation, based on an analysis of the pubic symphysis from different origins (dry bone vs. 3D reconstruction). This study involved a test sample of pubic symphyses, from which 3D reconstructions were obtained by scanning. We estimated the age of these individuals by comparing their pubic symphyses (derived from both dry bones and 3D reconstructions of the same individuals, and staged accordingly) against two distinct reference samples: the original Bone Reference Sample (BRS) by Brooks and Suchey, and a more recently published Virtual Reference Sample (VRS). The percentages of correct age estimates suggest no significant influence from the reference sample’s nature. For younger individuals, inaccuracy values reveal a lower age estimation error when utilizing the BRS reference sample. Furthermore, bias values show an overestimation. For older individuals (&gt; 40), the nature of the reference sample does not seem to improve the estimation. However, according to age groups, bias values seem to be less underestimated with the VRS reference sample. We can conclude that, regardless of the nature of the specimen, the BRS sample appears to improve age estimation for younger individuals. However, there seems to be no benefit to using a virtual sample for a 3D specimen. Furthermore, the age structure of the reference samples is an important factor to consider.</p>