Challenges in sample processing within radiocarbon dating and their impact in 14C-dates-as-data studies. (January 2020)

• Record Type: Journal Article
• Authors: Becerra-Valdivia, Lorena; Leal-Cervantes, Rodrigo; Wood, Rachel; Higham, Thomas
• Is Part Of: Journal of archaeological science. Volume 113(2020)
• Publication Date: 2020-01
• Subjects: Radiocarbon dating -- Summarized probability distributions -- Data analysis -- Archaeological science -- Archaeology; Archaeology -- Periodicals; Archéologie -- Périodiques; 930.1
• DOI: 10.1016/j.jas.2019.105043
• Languages: English
• ISSNs: 0305-4403

Abstract

Abstract: For decades, researchers have employed sets of radiocarbon dates to reconstruct trends in ancient human populations. The overarching assumption in this analysis is that the frequency of dates is proportional to the magnitude of past human activity. Thus, the distribution of summed or otherwise summarized dates is used to extrapolate population density and mobility patterns. There are, however, a number of underlying assumptions associated with this analysis that workers address to varying degrees and which, if false and not critically accounted for, will introduce bias, misrepresent the magnitude of activity, and ultimately prove misleading in archaeological interpretations. In this regard, research has so far mainly focused on correcting for the effects of time-dependent degradation of archaeological sites and constituent materials, calibration irregularities, and the efficacy of the statistical methods used. Assumptions directly related to sample processing in radiocarbon dating, however, are less discussed in ' 14 C-dates-as-data' analyses. It is, for example, assumed that all carbonaceous materials will yield sufficient, endogenous carbon for radiocarbon measurement. Yet sample failure in radiocarbon dating is common and contingent on, largely, deterministic factors such as post-depositional environment. Sets of radiocarbon dates analyzed, therefore, represent successful measurements independent of reliability. In this work, we discuss the biases introduced by challenges in radiocarbon processing and their impact on 14 C-dates-as-data studies. Highlights: The 14 C-dates-as-data method is critiqued from the perspective of radiocarbon specialists. Challenges in sample processing for radiocarbon dating impact a number of assumptions. To test the potential for bias, we recommend site-specific assessment of datasets. If a bias is unquantifiable or untestable, the sole use of this method is questioned.