That's a hot potato - we think a good research report should extensively document the issues identified in the research and separately provide a UX review exploring why the interface may be causing the issues the research surfaced, and suggest remedial approaches, for example a validated pattern.

Excellent point, and the interviewer should never assume and always probe, never with "You seem lost and frustrated, what's going on?" but neutrally "Can you tell me what's going for you right now?" and let the participant name themselves what's going on.

May we suggest using the term "signal" instead of "outcomes" for this case? A signal points to the problem but is not in itself the problem, as you point out. And this allows to reserve "outcomes" for business metrics, whatever they may be.

Great to see you here!