Search

Online supplement to the Çatalhöyük Volume 12 Chapter 9 [Examining Caprine Management and Cattle Domestication through Biometric Analyses at Çatalhöyük East (North and South Areas)]. Contains computational details of the mixture models (.Rdata file), R scripts, associated data sets, and Stan files to fully replicate the analyses and figures presented in the chapter.

EOC mantiene este repositorio dentro del Open Science Framework para mantener un DOI DOI 10.17605/OSF.IO/23BDT) y almacenar los archivos de datos, código y R Markdown necesarios para replicar la traduccion de este manuscrito al Castellano. Actualmente este trabajo esta en prensa con Advances in Archaeological Practice. Aquí los usuarios pueden producir versiones en PDF y HTML del manuscrito en Castellano para mayor comodidad. Sin embargo, tenga cuidado de que el código RMarkdown aún no esta completamente traducido. El PDF resultante todavía tiene partes escritas en inglés. Resumen Los arqueólogos utilizan con frecuencia las estadísticas de probabilidad y prueba de significación de hipótesis nula (NHST, por sus siglas en inglés) para evaluar qué tan bien se alinean los datos de encuestas, excavaciones o experimentos con sus hipótesis sobre el pasado. Cada vez más, los arqueólogos están adoptando estadísticas bayesianas para este propósito. Muchos arqueólogos están familiarizados con las estadísticas bayesianas, principalmente en la estimación de fechas de radiocarbono y la construcción de cronologías. Sin embargo, las estadísticas bayesianas tienen aplicaciones arqueológicas más amplias más allá de la cronología. El documento comienza contrastando brevemente los marcos estadísticos tradicionales y bayesianos. Luego, para guiar a los lectores a través de un ejemplo de inferencia bayesiana, el manuscrito presenta a "Monico", una cultura arqueológica ficticia, y al "arqueólogo famoso", una autoridad académica sobre Monico. El célebre arqueólogo ha excavado un yacimiento arqueológico de Monico rico en cultura material y compuesto por restos arqueofaunísticos. Luego, el manuscrito presenta el Teorema de Bayes y guía al lector a través de un análisis bayesiano elemental paso a paso de los restos de fauna como puntos de datos discretos para responder a una de las preguntas de investigación del famoso arqueólogo con respecto al comportamiento dietético de Monico. Después de que el arqueólogo responda su pregunta de investigación principal utilizando el teorema de Bayes, el manuscrito hace que el lector pase de los análisis bayesianos de puntos de datos discretos a análisis de distribuciones de datos. Todo en el contexto de la evaluación de hipótesis arqueológicas utilizando datos zooarqueológicos recuperados de múltiples sitios de Monico. Esta parte también guía a la audiencia a través de un ejemplo "formal" paso a paso de cómo los analistas contemporáneos aplican el teorema de Bayes a las distribuciones de datos para evaluar ideas científicas en arqueología.

Intra-tooth samples of enamel δ18O and δ13C isotopic values produce isotopic profiles that reflect seasonal fluctuations in temperature, precipitation, and dietary composition an animal experienced while its tooth was mineralizing. Archaeologists have interpreted trends found in isotopic profiles to estimate birth seasonality, to elucidate past management strategies (e.g., seasonal foddering), and to examine variability in management strategies related to other economic activities. These research aims rely on comparisons of isotopic profiles; Balasse, et al. (2012) developed parametric summaries of isotopic profiles that can help standardize comparisons across teeth. However, these summaries typically work best with an intensive intra-tooth sampling strategy, which is not always feasible due to preservation or budgetary constraints. This poster describes a Bayesian model for fitting a seasonal regression model to isotopic profiles of enamel δ18O and δ13C values in cattle teeth. Reasonable parameter estimates of the isotopic profile can be recovered from teeth sampled as few as 5-6 times, allowing for comparisons of relevant parameters within an assemblage. Comparisons of the profile parameters relate to different aspects of past herding systems and past ecology, such as seasonal dietary differences, strength of seasonality, and the spread of seasonal births.

The “Young Male Culling” model of early domestic animal herd management describes early herders as slaughtering excess males before they reached sexual maturity to maintain larger herds of adult females. Importantly for zooarchaeologists, this management system should produce distinctive patterns in the age and sex structure of the zooarchaeological assemblages, consisting of large-sized unfused specimens and smaller-sized fused specimens, particularly for elements with later-fusing epiphyseal ends (Zeder and Hesse 2000). While this pattern has been explored using log-standard index (LSI) values from archaeological sites to identify general biometric patterns, supplemented with mixture modeling (e.g., Arbuckle and Atici 2013), multilevel modeling of LSI values addresses biometric variation across different elements that can obscure patterns (Wolfhagen 2020). This poster combines multilevel LSI modeling with mixture modeling to estimate the age and sex structure of ten zooarchaeological assemblages of measured sheep (Ovis aries) remains from Neolithic sites in Anatolia. By using the parametric results of the biometric model to estimate the age and sex of unmeasured specimens, sex-specific age profiles can be constructed for different age stages. Comparisons of these sex-specific age profiles provides a framework to evaluate the probability that these assemblages suggest the presence of “young male culling” in Neolithic Anatolia. This project contains the code necessary to undertake the analyses in this poster. The poster was presented at the 2022 Society of American Archaeology conference in Chicago.

EOC maintains this repository within the Open Science Framework to keep a DOI (DOI 10.17605/OSF.IO/54F62) and store the R Markdown, code, and data files necessary to replicate this manuscript, currently under review with Advances in Archaeological Practice. Here users may produce PDF and HTML versions of the manuscript for convenience. ABSTRACT Archaeologists frequently use probability and null hypothesis significance testing (NHST) statistics to assess how well survey, excavation, or experiment data align with their hypotheses about the past. Increasingly, archaeologists are adopting Bayesian statistics for this purpose. Many archaeologists are familiar with Bayesian statistics, mainly in radiocarbon date estimation and chronology building. However, Bayesian statistics have broader archaeological applications beyond chronology. The paper begins by briefly contrasting traditional and Bayesian statistical frameworks. Then, to walk readers through an example of Bayesian inference, the manuscript introduces the "Monico," a fictitious archaeological culture, and the "famous archaeologist," an academic authority on the Monico. The famous archaeologist has excavated a Monico archaeological site rich in material culture and composed of archaeofaunal remains. The manuscript then introduces Bayes' Theorem and guides the reader through an elementary step-by-step Bayesian analysis of the faunal remains as discrete data points to answer one of the famous archaeologist's research questions regarding the Monico dietary behavior. After the archaeologist answers their primary research question using Bayes' theorem, the manuscript transitions the reader from Bayesian analyses of discrete data points to analyses of data distributions. All in the context of evaluating archaeological hypotheses using zooarchaeological data recovered from multiple Monico sites. This part also guides the audience through a step-by-step "formal" example of how contemporary analysts apply Bayes' theorem to data distributions to evaluate scientific ideas in archaeology

This project involves creating ordered logistic regression models for zooarchaeological survivorship analyses. Zooarchaeological data typically involve unknown ages-at-death, with elements assigned into ranked "age stages" that indirectly relate to biological age. The model developed in this project uses aoristic assignments to assign "probabilities of stage membership" to fragments, which are then fit using an ordered logistic regression.

This project includes the underlying data and analytical scripts necessary to replicate the figures in Hamilton, et al. (2021). Non-uniform tropical forest responses to the 'Columbian Exchange' in the Neotropics and Asia-Pacific. Nature Ecology & Evolution https://doi.org/10.1038/s41559-021-01474-4 The script produces individual GAMs for a single set of pollen core data over a specific period of time, including an uncertainty band corresponding to the parameter of interest. In the paper, the parameter of interest was NAP:AP (non-arboreal pollen:arboreal pollen) and other proxies for burning frequency.

This OSF Project contains the code necessary to reproduce the analyses in the preprint "Estimating the Age and Sex Composition of Faunal Assemblages with Bayesian Multilevel Mixture Models" by Jesse Wolfhagen, intended for submission to the Journal of Archaeological Method and Theory. It also contains code to apply the method to new faunal assemblages. For the most part, this code is held in the associated GitHub repository, though the Wiki here contains some of the same information and links to the original data sources of the data used in the manuscript and application code.

Environmental background and chronology