Guidelines for Statistical Usage in Age-Estimation Technics

Abstract

Most papers concerned with age estimation techniques contain procedural or statistical errors or omissions that hinder efforts to evaluate their accuracy and hence their worth. Eight common problems are cited: reconciliation of variability in counts of annuli within an individual, lack of known-age material, inadequate statistical treatment; and, for regression analysis, heteroscedasticity, non-linearity, confusion over the identity of the dependent variable, assessing accuracy of the technique and confidence limits. Recommendations are given to avoid these problems and to improve the quality of future age estimation literature.