Research · Kohavi, Tang and Xu, Cambridge University Press, 2020
Trustworthy Online Controlled Experiments: A Practical Guide to A/B Testing
experimentation practice
Supports the randomised holdout as the design that establishes cause by assignment, and names the conditions under which its reading can be trusted.
Read Kohavi, R., Tang, D. and Xu, Y. (2020). Trustworthy Online Controlled Experiments: A Practical Guide to A/B Testing. Cambridge: Cambridge University Press. · retrieved 2026-09-29
What it informs — 3
- enforcedCause is read by a declared design
- enforcedA value a plan fixes is a governed parameter
- enforcedA deliverable is a method the industry already writes
This source is what the model was built from. It does not itself commit the model to anything — a commitment does, and the ones it informs are named above.