Post-hoc same-system factual support review

The reviewer received the frozen source cases and all four response files, without condition labels or the answer key. It first considered three Q3 excerpts, then reviewed all 32 responses. Its final report follows verbatim.

Reviewed all 32 responses using only the five permitted files.

Clear unsupported additions:
- R01 / Q3: “Q3-A’s unresolved warning predates that correction.” The record’s creation date does not establish when its warning was written or revised.
- R04 / S3: “the catalog entry’s September 1 unresolved status.” September 1 dates the entry’s initial addition, not necessarily the unresolved-status text. This wording specifically dates the status.

Genuinely borderline:
- R03 / Q3: “The older setting warning.” Unsupported if “older” dates the warning’s content before the correction; defensible if it simply identifies the earlier-created setting record or describes its outdated status. I would not treat this as a clear error under the ordinary-reader instruction.
- R03 / S2: “rescheduled opening to October 15 and installation to October 10–14.” The source reschedules the opening but only authorizes those installation dates. Grammatically, the answer also claims installation was rescheduled, implying an earlier installation schedule. It could instead be read as compressed, slightly imprecise reporting of the approved schedule.

No other clear unsupported additions or material borderline phrasing found. In particular, the Q3 “supersedes” formulations reasonably describe which recorded status controls; they need not assert an undocumented edit or content-revision date.
