A Novel Approach to Selecting Samples for Radiocarbon Dating

The Bayesian statistical framework now offers an integrated approach to the interpretation of groups of related archaeological radiocarbon determinations. This increased interpretive power leads archaeologists to seek answers to ever more statistically complex questions. How does one select archaeological material and decide upon the number of radiocarbon samples to date in order to maximize gain in absolute chronological information and minimize cost? More specifically, how does one assess whether measurable improvements in dating accuracy can be achieved for an acceptable level of resource investment? We choose a risk analysis approach which allows us to balance cost against information gain and to select which samples to date on this basis. Our approach is set within the Bayesian framework and can thus be used in conjunction with other, previously published, radiocarbon calibration procedures.