Reproducibility Package: Residual-First Benchmark for Model Adequacy Under Declared Measurement Uncertainty

This repository provides the complete reproducibility package for the paper: “A Residual-First Benchmark for Model Adequacy Under Declared Measurement Uncertainty.” Contents include: - Full calibration dataset (n = 48)- Declared uncertainty model- Model-fitting scripts (Python)- Residual and normalized residual calculations- Declared modeling sensitivity S(x)- Simulation study (300 runs across three scenarios)- Summary tables and figures All results reported in the manuscript can be reproduced using a single execution script. The package enables full verification of Example B and the simulation study, ensuring transparency, auditability, and alignment with GUM-based uncertainty evaluation practices.