On the accuracy of estimating living stature from skeletal length in the grave and by linear regression


  • H. C. Petersen

    Corresponding author
    1. Department of Statistics and Anthropological Data Base, University of Southern Denmark, Denmark
    • Department of Statistics, University of Southern Denmark, J. B. Winsløwsvej 9B, DK-5000 Odense C, Denmark.
    Search for more papers by this author


This study evaluates a method for obtaining stature estimates for populations represented by skeletal material, with individuals buried in a supine position. During the excavation of a Danish mediaeval cemetery, in situ skeletal length in the grave was measured from a point above the cranial point farthest from the body to the most distal point of the talus. The measurement was made with a folding rule placed on the sagittal midline of the skeleton, allowed to follow any curvature of the skeleton in situ. In the laboratory, stature was reconstructed anatomically, and this stature was regarded as an accurate estimate of living stature. Stature was also reconstructed from femur length by two linear regression procedures: 1) by sample and sex specific formulae, employing a leave-one-out approach, and 2) by sex wise formulae for Euro-Americans from Trotter & Gleser (1952, American Journal of Physical Anthropology10: 463–514). Skeletal length in the grave and the two stature estimates based on linear regression were compared to anatomically reconstructed stature. Skeletal length in the grave estimated anatomically reconstructed stature with practically no bias (95% CI: −1.3–1.5 cm). Sample specific regression formulae estimated anatomically reconstructed stature also with no bias (95% CI: −1.2–1.1 cm). In contrast, statures calculated from Trotter & Gleser's regression formulae estimated anatomically reconstructed stature with a bias of about 4 cm (95% CI: 3.3–5.0 cm). Estimates of stature variance were biased for all three estimation procedures. However, for samples of adults, an adjusted variance estimate can be obtained by subtracting 8.7 cm2 from the variance obtained from skeletal lengths in the grave. It is recommended to measure skeletal length in the grave whenever possible, and use this measurement for estimating statures for prehistoric and early historic populations. Copyright © 2005 John Wiley & Sons, Ltd.