Combining archaeological and radiocarbon information: a Bayesian approach to calibration

A recent and significant improvement in radiocarbon dating has been the increased ability of the radiocarbon laboratories to provide results combining precision with accuracy. This improvement has been accompanied by increasing recognition that the information must be expressed on the calendar, rather than on the radiocarbon, time-scale. Despite the attempts of Ottaway (1987) and Pearson (1987), archaeologists are not sufficiently aware of the statistical problems involved in the transformation from one scale to the other: ‘Some of the trouble lies in the ignorance of radiocarbon consumers; the many attempts to educate them can have only limited success when radiocarbon study depends on statistical concepts and methods far beyond the average archaeologist’s innumerate grasp’ (Chippindale 1990: 203).