Method and Theory in Deciding Identity of Skeletonized Human Remains

Abstract

All identifications in forensic science take the logical form of λ (lambda) = np, where λ = predicted number of occurrences of a combination of identifying traits, p = proportion of general population with that combination of traits, and n = size of population from which the particular occurrence was drawn. In human skeletal remains identification where it is desirable to reduce ‘n’ and/or ‘p’ maximally, the ‘missing persons’ list from CPIC effectively defines ‘n’ as a category of individuals whose age, sex, race, stature range, death interval range and geographic locale match the remains. The proportion (p) of the population sharing the particular combination of identifying traits seen on the remains can be determined either from purpose-specific epidemiological sampling or by multiplication of each trait's separate proportion (probability), as provided in the scientific literature, to find the probability of all those traits occurring jointly. The latter approach is valid only when the occurrence of one trait is independent of the occurrence of other traits in the joint probability calculation. The magnitude of interdependence can be determined and reduced by collecting prevalence data for each trait in terms of the major biological variables on which they depend; e.g., age + sex + race + socioeconomic group. Application of the method to a particular case of found human remains provides an estimate of the probability of identity of.95. The quantitative approach to human skeletal remains identification advocated here, while imperfect because of factors such as residual trait interdependence, provides a replicable, objective guide to deciding identity which can be subject to scrutiny by the courts, lawyers and coroners.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 10 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.