DSP: A probabilistic approach to sex estimation free from population specificity using innominate measurements

Abstract This chapter discusses the DSP (Diagnose Sexuelle Probabiliste), a probabilistic method for sex estimation based on a dataset of 2040 innominates corresponding to the variability of human species in time and space. An independent sample was used for the validation of reliability of DSP2, and confirmed that sexual dimorphism of the pelvic region is not population-specific. The method offers some flexibility in the use of variables (a minimum of 4 out of 10 measures are needed to estimate a sex), which minimizes the problem of bone preservation. After its first publication in 2005, the DSP, which is based on Fisher's linear discriminant analysis, is now available in a free and cross-platform software. DSP2, made available in 2017, is one of the best tools for sex estimation both for bioarchaeology and forensic anthropology use, where the quantification of results is a major benefit.