Introduction: This paper expands on recent research by Milella et al. (2015) on the classification of documented occupations in the studies of entheseal changes (EC) addressing the following questions: can EC variability be used as classification criterion of occupations and, by proxy, assess physical activity? Which analytical protocols are more suitable when trying to control for the effect of age? Materials and Methods: Both bivariate (linear regression - LR) and multivariate (nonlinear principal component analysis – NLP and hierarchical cluster analysis - HCA) statistical protocols were used on a dataset (n=372, male individuals, 20-88 y.o.) of postcranial entheses from three contemporary Italian and Portuguese identified skeletal collections (ISC). The study focused on entheseal robusticity (ER) (i.e. relative surface roughness), on laterality separately and together, and on levels of asymmetry. Variance in robusticity was calculated after NLP for each occupation. Age-at-death was controlled using the residuals obtained after regressing each dimension obtained from NLPC with age and exploring possible groupings of occupation by means of HCA. Results: Robusticity variance showed a complex picture characterized by an overall separation between “high variable” and “low variable” occupations, which in some cases is consistent with specific exposures to distinct biomechanical patterns. Clusters of occupations based on residuals revealed a subdivision in three main classes as observed in Milella et al. (2015). The occurrence of problematic associations traces a more complex picture in which inherent sample biases, incomplete documentation and poor reliability of the EC investigated may have played their part. Discussion and Conclusions: Our results demonstrate that: (1) variability of ER is an interesting approach when studying subdivisions of documented occupation; (2) regression residuals are useful for controlling the effect of age on ER; and (3) the correlation between EC and specific behavioural patterns may be problematic (i.e. lack of homogeneity and documentation characterizing ISC).
(Mis)informed? Identified skeletal collections, occupation, age at death and the reconstruction of physical activity
Milella M;
2016-01-01
Abstract
Introduction: This paper expands on recent research by Milella et al. (2015) on the classification of documented occupations in the studies of entheseal changes (EC) addressing the following questions: can EC variability be used as classification criterion of occupations and, by proxy, assess physical activity? Which analytical protocols are more suitable when trying to control for the effect of age? Materials and Methods: Both bivariate (linear regression - LR) and multivariate (nonlinear principal component analysis – NLP and hierarchical cluster analysis - HCA) statistical protocols were used on a dataset (n=372, male individuals, 20-88 y.o.) of postcranial entheses from three contemporary Italian and Portuguese identified skeletal collections (ISC). The study focused on entheseal robusticity (ER) (i.e. relative surface roughness), on laterality separately and together, and on levels of asymmetry. Variance in robusticity was calculated after NLP for each occupation. Age-at-death was controlled using the residuals obtained after regressing each dimension obtained from NLPC with age and exploring possible groupings of occupation by means of HCA. Results: Robusticity variance showed a complex picture characterized by an overall separation between “high variable” and “low variable” occupations, which in some cases is consistent with specific exposures to distinct biomechanical patterns. Clusters of occupations based on residuals revealed a subdivision in three main classes as observed in Milella et al. (2015). The occurrence of problematic associations traces a more complex picture in which inherent sample biases, incomplete documentation and poor reliability of the EC investigated may have played their part. Discussion and Conclusions: Our results demonstrate that: (1) variability of ER is an interesting approach when studying subdivisions of documented occupation; (2) regression residuals are useful for controlling the effect of age on ER; and (3) the correlation between EC and specific behavioural patterns may be problematic (i.e. lack of homogeneity and documentation characterizing ISC).I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
