How Many Dates Do I Need?

Abstract

As the use of large-scale radiocarbon datasets becomes more common and applications of Bayesian chronological modeling become a standard aspect of archaeological practice, it is imperative that we grow a community of both effective users and consumers. Indeed, research proposals and publications now routinely employ Bayesian chronological modeling to estimate age ranges such as statistically informed starts, ends, and spans of archaeological phenomena. Although advances in interpretive techniques have been widely adopted, sampling strategies and determinations of appropriate sample sizes for radiocarbon data remain generally underdeveloped. As chronological models are only as robust as the information we feed into them, formal approaches to assessing the validity of model criteria and the appropriate number of radiocarbon dates deserve attention. In this article, through a series of commonly encountered scenarios, we present easy-to-follow instructions for running simulations that should be used to inform the design and construction of chronological models.

The Symbiosis of Lichenometry and Radiocarbon Dating: A Bayesian Chronology of Alpine Hunting in Colorado’s Southern Rocky Mountains, Usa

“…Samples sizes increased by 10 for each model iteration. I accepted the results of the simulation tests when sequential iterations produced start and end boundaries that exhibited diminishing returns for improving precision (Holland-Lulewicz and Ritchison 2021:276–277). Model code and simulation code are provided in Supplementary Material .…”

“…In Model B I combined several outlier models (Bronk Ramsey 2009b; Dee and Bronk Ramsey 2014), including an “old wood” model for dated samples with inbuilt age issues (bulk charcoal, other large wood fragments, or unknown samples) and a general outlier model for other dates (lichenometric, bone collagen, charred twigs and needles). A series of simulations were then used to examine the sensitivity of the Model B output (Griffiths 2014; Holland-Lulewicz and Ritchison 2021). I tested for model reproducibility by randomly sampling a range (population mean (μ) ± 1σ) within the marginal posteriors of dated events from Model B.…”

“…I applied additional simulations to the existing set of dates from Model B (with flagged outliers manually removed) to reveal the effect of increasing sample sizes of high-precision 14 C dates. This method examines how robust the model output is to the addition of new high-quality dates, and in effect, reveals the minimum number of new dates needed to increase precision for estimates of the start and end boundaries for the game drive phase (Holland-Lulewicz and Ritchison 2021). For this test I generated random sample sets of simulated 14 C dates and constrained the date ranges for expected calendar dates based on the results of Model B. I provided a constant error of ± 20 years for each simulated 14 C date, assuming lab errors with high precision.…”

Statistical Inference of Prehistoric Demography from Frequency Distributions of Radiocarbon Dates: A Review and a Guide for the Perplexed

“…Carry-out tactical models and what-if experiments. Tactical models (Crema, 2018;Lake, 2014;Orton, 1973) and what-if experiments (Buck & Meson, 2015;Hinz, 2020;Holland-Lulewicz & Ritchison, 2021) are simulation techniques consisting of generating, in silico, artificial archaeological data under known conditions to determine the robustness of analytical techniques, explore the impact of particular biases, or estimate necessary sample sizes and guide data collection. These are powerful yet relatively underutilised tools that can enormously help in any statistical analysis.…”

Statistical inference of prehistoric demography from frequency distributions of radiocarbon dates: a review and a guide for the perplexed

“…models(Orton 1973, Lake 2014, Crema 2018) and what-if experiments Meson 2015, Holland- Lulewicz andRitchison 2021) are simulation techniques consisting of generating, in silico, artificial archaeological data under known conditions to determine the robustness of analytical techniques, explore the impact of particular biases, or estimate necessary sample sizes and guide data collection. These are powerful, yet relatively underutilised tools, that can enormously help in any statistical analysis.…”