Part IV · Chapter 43Version 1.1 · 5 November 202631 min read
Small Samples and Sparse Subgroups
Key message
Small groups are where fairness harms are most likely to hide and hardest to measure; assessments must report them with exact or score-based intervals, correct for multiplicity and estimator bias, borrow strength transparently and never drop them silently.
Five chapters or files are free. This one would be your sixth.
Leave an email address and every chapter and every PDF of the compendium opens, in this browser, for a year. Nothing else is asked, and whether we may contact you is your choice.