Statistical analysis is fundamental to science, and support decision-making in the public and private sectors. While statistical algorithms have advanced rapidly, software support and modelers' knowledge have failed to keep pace. This mismatch has resulted in suboptimal statistical workflows contributing to the reproducibility crisis in science. My team has significantly advanced the development of both theory and methods for safe iterative model building. Nonetheless, they have not yet been widely adopted by users. This is because implementing statistical methods in software is far from trivial, requiring careful consideration of mathematical, algorithmic, and computer limitations, implementation details, and human cognitive constraints. This project will significantly improve open-source software tools for safe iterative Bayesian model building, with immediate high impact, enhancing the efficiency, flexibility, and trustworthiness of model building.

2026

2027