Creating Reproducible Sum Calibrations for Archaeological Demographic Assessments. The R Package ‘OxcAAR’

Abstract

The use of sum-calibrated 14C data as a demographic indicator is a controversial field of archaeological methodology. It requires the largest possible number of source data, which in themselves only have a weak coupling to the phenomenon of investigation (popula- tion) and are influenced by a wide range of different biases. This makes it particularly important to disclose the data basis as well as the methodology of data processing: A high degree of transparency can serve (not only) in this case to establish a basis of trust from the very beginning. This is all the more relevant as Monte Carlo simulations already have established themselves as a gold standard for assessing the robustness of population size extrapolations. We (the working group ISAAKiel) have developed the R package oxcAAR to facilitate accessibility to the methodology, the stand- ardisation of the procedure and thus the comparability of various studies. It is designed to obtain sum calibrations from within the statistical framework R (which is excellently applicable for reproducible research and Monte Carlo simulation) by means of a gener- ally accepted calibration standard (OxCal), which serves as a basis for the reconstruction of demographic developments. We would like to take this opportunity to familiarize the listeners with the underlying issues and the solutions our workflow provides. This will be done with the help of another R package developed by us: c14bazAAR, a database crawler for 14C data. In the lecture, we will go through our workflow in an exemplary manner and also discuss the general pitfalls involved in sum calibration and how to avoid them.