Survey Cognitive Pretest

Act as a Principal UX Researcher who runs cognitive interviews on questionnaires before fielding.

I have a draft survey I need to pretest with a small set of people who match the audience.

Survey purpose / decision this will inform: [what will change based on results?]
Target audience: [who should take this]
Draft survey (questions, scales, instructions, skip logic):
[Paste the full instrument]

Please produce:

1. Pretest protocol (20-25 minutes)
   - Intro script (neutral; do not explain what each question "means")
   - How to have them think aloud while answering
   - When to probe vs. stay quiet
   - Closing: what felt long, confusing, or skippable

2. Item-by-item watch list
   For each question, note:
   - What a respondent might think the question is asking
   - Likely misread, jargon, or double meaning
   - Scale problems (labels, midpoint, "N/A" missing)
   - Skip-logic traps

3. Five probes to use live
   Short, reusable lines (e.g., "What does [word] mean to you?" "How would you answer if you had never used the product?").

4. Revised survey
   Rewrite only items that fail a think-aloud. Keep validated scales (NPS, SUS, UMUX-Lite) intact unless surrounding instructions cause the failure. Mark every change.

5. Fielding go / no-go
   Estimated completion time, drop-off risks, and whether this is ready for a 5-10 person silent pilot or still needs another wording pass.

Do not invent pilot results. This is a protocol and a rewrite, not fake data.

Follow-ups: "Rewrite for non-native English readers at about 8th-grade level." or "Turn the watch list into a one-page moderator crib sheet."