atrix-Informed Correction and Exclusion Protocol for Biogenic Radiocarbon Dating: Demo Dataset and Code (v1.3)

This archive provides the compact demo dataset and minimal code for the matrix-informed correction and exclusion protocol for biogenic radiocarbon dating described in the paper “Matrix-Informed Correction and Exclusion Protocol for Biogenic Radiocarbon Dating: A Physical–Chemical Framework for Plant, Bone, and Carbonate Samples”. The goal of the package is to make the protocol fully reproducible and to serve as a teaching resource for matrix-aware quality control in radiocarbon work. The data/raw/ directory contains synthetic but realistic CSV tables that emulate standard laboratory spreadsheets: field metadata, radiocarbon measurements, QC metrics (collagen yield, C/N, FTIR IRSF, calcite %), QC notes, and site-level environmental covariates. The data/processed/ directory holds analysis-ready tables derived from these inputs, including a consolidated pair-level dataset (dataset_radiocarbon_pairs.csv), QC summaries, and three small literature-based case-study tables: freshwater shell–charcoal pairs from Elk Hills (CA-KER-116), Vindija/Zlatý kůň bone pretreatments (legacy collagen vs UF vs HYP), and leaf-wax vs macrofossil ages from Lake Chichancanab. The code/ directory contains minimal R and Stan skeletons (run_all.R and model_hierarchical.stan plus helper scripts) that show how to recreate the processed tables, fit a simple hierarchical matrix–site model, and export summary CSVs in results/summaries/. The docs/ and materials/ folders provide documentation stubs (CODEBOOK, WORKFLOW, REPRODUCTION_CHECKLIST) and a reporting-checklist template for matrix-informed radiocarbon datasets. All numbers used in the figures of the accompanying paper can be recomputed directly from the CSV files in data/processed/ and results/summaries/; figure PDFs themselves are not included in this archive. The contents are released as an open reference implementation for the matrix-informed protocol and may be freely reused under the Creative Commons Attribution 4.0 International (CC BY 4.0) license.