When analyzing human adult skeletal remains, it is often difficult to decide whether a single aging method will give a more reliable age estimation than a combination of methods. This study evaluates four macroscopic indicators for age estimation on 218 American White and Black individuals, ranging in age from 25 to 90 years of age, from the Terry collection. Individuals in the sample were selected to have a balanced race, sex, and age distribution. The following aging methods were applied to each skeleton by one experienced observer: the Suchey-Brooks (SB) pubic symphysis method, the Lovejoy auricular surface method, the monoradicular teeth Lamendin (LM) method, and the Işcan (IC) method for fourth ribs. The statistical study involved the evaluation of inaccuracy and bias (based on median age) for each age indicator and the combination of methods using Principal component analysis (PCA). Analysis was performed on the entire sample, then by race, then sex, and then age group (25-40 years, 41-60 years, and >60 years). PCA was the most accurate method for both racial groups when all age groups are analyzed together. When the sample was divided into age groups, SB was the most accurate for young adults (25-40 years) and LM was the most accurate for middle adults (41-60 years). After the age of 60, all methods are highly inaccurate, although IC gives the lowest inaccuracy. As regards bias, the study highlights the tendency of all methods to overestimate the age of young individuals and to underestimate in the older age group. No single skeletal indicator of age at death is ever likely to reflect accurately the many factors that accumulate with chronological age. In fact, one must use as many dental and skeletal indicators as possible. However, in order to maximize the potential of each method, in the final evaluation one should consider mainly the method or methods that have a higher accuracy for a particular age range.

No abstract

International audienceCrop models are important tools in agronomic research, a major use being to make predictions. A proper parameter estimation method is necessary to ensure accurate predictions. Until now studies have focused on the application of a particular estimation method and few comparisons of different methods are available. In this paper, we compare several parameter estimation methods, related, on the one hand, to model selection, and on the other, to ridge regression based on an analogy to a Bayesian approach. The different methods are applied to a simplified crop model derived from the STICS model, using simulated data. The criteria for comparison are prediction error and errors in the parameter estimates. Among the methods of model comparison a version of the Schwarz criterion, corrected for small samples and with maximum and minimum bounds for each parameter, is the preferred method. Ridge regression is found to be superior to this best method of model selection

Lamendin and colleagues (1992) proposed to assess age-at-death from root translucency and periodontosis. Several teeth from the same individual were included in their study. In our work, we evaluate the reliability of their formulas without introducing this bias. Our sample was constituted of 214 individuals (114 males and 100 women) selected from the Terry collection (U.S.A.). The R(2) between age and both indicators is equal to 0.33 and 0.08 (p < 0.001). Based on linear regression, the mean of standard error for individual age prediction was equal to 13.67 years, and the mean interval length is equal to 53.89 years. Multinomial logistic regression and Bayesian approach failed to give satisfactory results when classifying the individuals in age categories. Therefore, the use of root translucency and periodontosis may lead to incorrect age-at-death assessment, and it is thus necessary to complement this approach with other techniques to assess age-at-death.

No abstract