The domain of validity

This archive contains a technical report, its appendix, eight figures, and a set of standalonechecks. The checks are NumPy-only and can be run without the data (29 analytic self-tests; a paritytest against the original pipeline is skipped when that pipeline is absent). The cohort statisticsare not reproducible from this archive: the five source datasets are third-party and are notredistributable. The archive is offered so that the method can be examined and the checks can berun, not as a claim of full computational reproducibility.Text and figures: CC BY 4.0. Code in checks/: MIT.