D1 | Reviewer A codes R4 as search failure; reviewer B codes it as invitation/access permissions. Resolve by source evidence and separate observation from hypothesis, not a vote.
D2 | R1 may receive both large-file performance and completion-feedback codes. Do not invent an unmeasured delay.
N1 | participant U12 | additional interview | The v2.1 help page showed an older invitation screen. I asked an administrator which button to use on the new screen. I did not test whether my permissions were wrong.
D3 | Report record counts, unique participants, independent completion and counterexamples separately. A suggested support investigation is not research already performed.