This paper investigates the value for forensic anthropology of craniometric data in assessing population affinity. It finds that generally speaking cranial measurements do not contain the information to directly make a positive match for a skull's population affinity. Rather, cranial measurements should be thought of as containing information that allows for the elimination of any population affinity for the skull which would be a mismatch. A minimum of 13 measurements is required to capture enough information to be confident that the eliminated population affinities are indeed the mismatches. In addition, if a reasonably sized sample of crania from the same population is available for analysis, the affinities of the sampled population can be reliably assessed using the methodology outlined in this paper.