Randomisation, blinding limits, and the logging interval. The method is unglamorous, which is precisely why the finding is worth reading.
1 branch open from this page.
A section is the smallest unit that still argues something. If it can be removed without changing the claim, it does not belong in the dossier.
This level sits inside The twelve-week evening-consistency trial. It is written to be read on its own, but it assumes the claim made one level up. The dossier only works when the levels agree with each other.
If the measure can be gamed, the finding will be a finding about gaming.
The standard is the same at every depth: state the claim, show the evidence, name the limit, and link the source. Anything that cannot meet that standard is moved to a field note and marked as provisional.
Sections are written to be quoted. Each one carries its own claim, evidence, and limit.