Variation within physical and digital craniometrics

Metric assessment of human crania can provide forensic practitioners and anthropological researchers with information on an individual's sex and biogeographical ancestry. However, metric methods rely on the ability of users to remain consistent with themselves and others, with any error in the data rendering conclusions invalid. Digital anthropology is a growing sub-field where human remains are digitised using a growing range of methods and technologies. These models have the potential to boost research collaboration and public engagement. However, not all of these digitisation methods have been examined critically to explore the veracity of their use within a research environment. There has also been limited research into the application of digital anthropology to craniometric analysis. This study examined the intra- and inter- observer variation of seven participants taking physical measurements from a human cranial cast with an associated set of reference values. The same measurements were also taken from three digital models of the cranial cast which were created using digital photogrammetry and laser scanning. This data was then compared to the reference values and the physical measurements taken by the lead author. This study found that there was excellent statistical agreement between the reference values and the measurements taken from the cranial cast, both physical and digital. However, the participants still exhibited variation within a range of -18mm and +30mm from the reference values. MANOVA tests showed between-subject effects on nine measurements across the participant data, and 12 measurements between the digital models. However, there is little consistency between this study and the anthropological literature as to which measurements are most prone to between-subject effects. Despite the excellent agreement shown between the reference values and the digital models this study raises a number of methodological questions regarding inter-observer error and the varying levels of data processing present in different digitisation methods.