Growth deficit in survivors and non‐survivors: Biological mortality bias in subadult skeletal samples

Linear growth of the long bones has and continues to be used by osteologists as a non‐specific indicator of health. Studies of past populations use cross‐sectional mortality samples to construct human skeletal growth profiles as a means of inferring the overall health of the population from which the skeletal series was drawn. Although all studies recognize that the growth profiles derived from skeletal samples may not be representative of the “true” path followed by the children who survived to adulthood, only recently have there been attempts to quantitatively examine this problem. This paper addresses the issue of biological mortality bias in subadult skeletal samples, focusing specifically on reduced or retarded linear growth. We begin with an overview of the general opinion on the subject within the osteological literature. A review of the child survival literature examines the interrelationships between malnutrition, morbidity, mortality, and linear growth, considering both indirect and direct studies of survivor–non‐survivor growth data. This literature suggests that mortality samples are potentially biased, with linear growth of survivors often greater than that of non‐survivors. The issues of stunting (the process of becoming small) and stunted growth (being small) are discussed with respect to biological mortality bias. It is concluded, that while the potential for such a bias exists within subadult skeletal collections, the effects are likely to be small at the aggregate level and error introduced by other methodological considerations (ageing, unknown sex, sample size, preservation, quality of excavation) is likely to outweigh any such error in interpretations of past population health. © 1993 Wiley‐Liss, Inc.