1Laboratory of Survival & Longevity, Max Planck Institute for Demographic Research, 2Anthropological Database Odense University, University of Southern Denmark, 3Laboratory of Statistical Demography, Max Planck Institute for Demographic Research, 4Department of Anthropology, Pennsylvania State University
Thursday All day, Plaza Level
Over the last 250 years, the mean length of the human lifespan increased in an approximately linear fashion. However, all information has been obtained through the analysis of written sources. To gain a better understanding about the processes that shape human mortality, especially in old ages, precise information about age-at-death derived from biological remains is necessary.
Most methods of estimating age from skeletons result in open age intervals, which often start as early as 50 or 60 years of age, and lack accuracy and precision. Calibrated Expert Inference (CEI) solves these methodological problems by combining the knowledge about new -as well as some established- osteological indicators with statistical analysis to derive unbiased estimates of age at death. A single age indicator, the so-called “expert age”, is processed statistically to derive an individual’s age estimate. By using a non-parametric regression technique, which allows for interval-censored data, there is no need to rely on linearity between the indicator and actual age, nor on homoscedastic error distributions. A maximum-likelihood procedure is used to estimate the population mortality pattern before the individual age is calculated by applying Bayes’ Theorem.
We will present the osteological indicators that are used to obtain the initial “expert age” and show the first results where CEI has been applied to populations from historical Scandinavia. These show that old ages can be identified with precision in skeletal material, and therefore it is possible to draw conclusions about the development of longevity in the past.