Indirect estimation of age: applications in paleo-, medieval and contemporary demography

La plupart des méthodes employées en paléodémographie pour estimer la structure par âges d’une population inhumée sont entachées de biais méthodologiques les rendant inutilisables. Récemment Caussinus et Courgeau (2011) ont élaboré une méthode reposant sur une procédure d’inférence bayesienne, censée pouvoir être utilisée avec n’importe quel indicateur biologique, pour estimer avec fiabilité la structure d’une population d’âges inconnus. Nous faisons donc l’essai sur un nouvel indicateur: le cément dentaire. La cémentochronologie est une méthode d’estimation de l’âge individuel reposant sur le décompte des dépôts annuels de cément acellulaire autour des racines dentaires. Cette méthode a été mise au point par les zoologues pour déterminer l’âge d’espèces animales, avec de bons taux de corrélation. Elle est actuellement appliquée à l’homme et les premières études publiées s’accordent à trouver cet indicateur très prometteur. Le travail que nous présentons repose, dans un premier temps, sur la mise au point d’une procédure pour passer d’une lecture directe d’un âge individuel à une estimation de la structure par âge d’un ensemble d’individus. Une fois élaborée la table de contingence nécessaire à l’implémentation de la méthode PIB, nous testerons les résultats sur une ou deux populations d’âges connus, prises hors échantillon

A recent study focusing on Italian Early and Late Middle Ages using archaeological data shows that the ratio between male and female individuals aged 20+ was clearly disproportionate in favour of the former (169 men per 100 women). Such unbalanced sex ratio is not observed neither in Italian cemeteries during Roman times or Late Antiquity nor in medieval necropolises of France, Hungary and Germany. The first aim of this paper is to demonstrate that this disproportionate sex ratio mirrors a higher infant and youth female mortality in the Italian Middle Ages, rather than statistical artefacts or selection problems concerning data on the necropolises compared those on the general population. The second aim is to disentangle the causes of such peculiarities. The (albeit scarce) written sources support the idea that the transition from Roman to medieval times in Italy was characterized by a weakening of the pater familias and – more generally – the family of origin’s control over women (throughout their lives), in part due to a widespread affirmation of exogamous marriage. This shift may have been particularly drastic in Italy – the heart of the Roman familial system – compared to Central-Eastern Europe.

Age of the couples is one of the most basic and important information collected in almost all demographic and health surveys. This particular study makes an attempt in exploring the extent of age difference and the change over time among the couples in India. It also analyses the impact of other socio-economic, demographic and cultural factors in explaining the variation in age difference among the couples. The study analyses the age difference among the husbands and wives on the basis of the reporting of the age of the couples by the wives by using a nationally representative sample of the three rounds of National Family Health Survey conducted at three different points of time (1991-92, 1998-99 and 2005-06). The present study uses the data of 257629 (combined data for NFHS-1, NFHS-2 and NFHS-3) currently married women who married only once. It is found that in India the age difference among the couples has decreased over the period of time from 7 years during 1956-60 to 5.2 years during 2001-05.

Indirect estimation of age of human skeletons and the proper methods to attain it are basic to studies of past populations. One of the basic questions is how the life spans and pattern of adult mortality may differ from the contemporary situation. In the 10 years since the publication of the Rostock Manifesto, which famously set forth aims for paleodemography, recent investigation in improving age estimation, especially for older adults, has increased. New methods, such as transition analysis, try to meet the Rostock goal of better estimations of the probability of age given the stages of age indicator(s), of a target population of unknown ages at death. However, some of these seem to result in unrealistic old age estimates, which are generally just brushed aside as quirks of the methods used. However, if results are unreasonable at older ages, why should they be better at slightly younger adult ages? What might be reasonable life spans in preindustrial, non-Western populations? We investigate this problem with skeletal samples from two neighborhoods in Teotihuacan, Mexico. Investigating how differences in the Precolumbian urban environment might have influenced adult mortality and adult lifespan should prove a corrective to pure model-based estimations

The analysis of demographic behavior with respect to age has been evaluated with various methodologies. With the purpose of comparing differences between traditional or classic aging methods and the transition analysis method proposed by Boldsen and colleagues (2002), we reevaluated age estimates in the San Gregorio Atlapulco skeletal assemblage, an osteological collection that dates to the end of the prehispanic period and that we consider suitable for paleodemographic studies due to the composition of the sample and its large size. The applied methodology resulted in substantial differences. It overestimated the skeletal age of the oldest adults, in comparison with previous estimations derived from traditional aging methods from physical anthropology, pushing the maximum age in the group to 100 years or more. These results will be discussed with respect to the principles of anthropological demography and previous studies of health in this population.