The 81st Annual Meeting of the American Association of Physical Anthropologists (2012)

Meta-analysis of forensic stature estimation


Behavioral Sciences, University of Michigan-Dearborn

Thursday All day, Plaza Level Add to calendar

Accurate stature estimation relies on population, age, sex and generation specific formulae due to varying body proportions worldwide. The goal of this research project is a global meta-analysis of stature estimation. A survey of more than sixty recent publications reveals the current state of stature estimation; what populations have been studied and what populations require data collection and/or analysis. This paper also reviews the estimated number of missing persons for different regions.

Of the publications reviewed, the most accurate predictors of stature continue to be the femur and tibia. The coefficient of correlation for which ranges from 0.82 to 0.93 and SEE of 2.4-4.0 for both males and females. Thirty of the studies compare the long bones of the upper and lower limbs. Fifteen of the studies compare the measurements of the hand or feet (fleshed, radiographic, tarsals or metacarpals/metatarsals). The vertebrae are the next most commonly compared elements to stature. A handful of the studies look at measurements of the cranium, face or dentition, but with little success (r<.06, SEE >5.0). The cranial sutures had the lowest coefficient of correlation (.09-.363). Multiple regression yields higher results than simple linear regression. Stature formulae for males tend to have higher correlations than those for females.

The regions that have the highest number of stature formulae for the populations represented are North America, Eastern Europe, West and South Asia and Southern Africa. Areas that would benefit the most from new stature estimation include South America, North and Central Africa and Southeast Asia.

comments powered by Disqus