[Heidelberg Centenarian Study: theoretical and methodological principles for social science research of the oldest old].

This paper addresses two important issues of oldest old research. First, we try to answer the question whether differences between extremely old persons are based on enduring aging processes of physical, social and psychological functions, or whether they are the result of selective survival weeding out the frail. Both trajectories predict low variability, but on different levels. Second, we investigate to which extent the variability of extremely old persons can be assessed by traditional psychosocial instruments. One hundred and fifty six centenarians were contacted, of whom 91 centenarians and 86 proxies participated in this study. The distribution of the cognitive status reveals a high amount of variability. In contrast, with respect to functional capacity 83% receive support from the German Statutory Insurance System for Long-Term Care. The higher variability in the psychological variable points to adaptive potential. We assume that in extreme old age various resources are used to maintain physical functioning above a critical threshold.