/verified-infographic-generator

Verification Report — <slug>

Date: YYYY-MM-DD Poster topic: <one-line description> Claim Plan: 01-claim-plan.yaml

This report is the user checkpoint artifact. It shows every claim from the Claim Plan, what the web search turned up, and how each claim was classified. No layout or image generation happens until the user approves this report.

Summary

Abort threshold check: If rejected claims > 20% of planned claims, halt and reshape the poster with the user.

Claim-by-Claim Results

claim_id: example_claim_1

• Subject: well-being in biophilic offices
• Classification: VERIFIED
• Final value: +15%
• Final phrasing: "Self-reported well-being increase"
• Source: Human Spaces / Interface (2015)
• Authors: Cooper, C. et al.
• URL: https://www.interface.com/...
• Quoted passage: "employees who work in environments with natural elements report a 15% higher level of well-being"
• Search queries used:
  • "Human Spaces Global Report 2015 biophilic design productivity well-being"
  • "Interface 2015 biophilic office survey 7600 workers"

claim_id: example_claim_2

• Subject: cortisol response in natural environments
• Classification: QUALITATIVE-ONLY
• Final value: "Significant reduction" (no percentage)
• Final phrasing: "Nature exposure reduces cortisol (Antonelli et al., 2019 meta-analysis)"
• Source: Antonelli, Barbieri & Donelli (2019). International Journal of Biometeorology.
• URL: https://pubmed.ncbi.nlm.nih.gov/31001682/
• Quoted passage: "Forest bathing can significantly influence cortisol levels on a short term in such a way as to reduce stress."
• Why not quantified: Meta-analysis reports significant effect sizes but no single clean percentage appropriate for a hero number.

claim_id: example_claim_3

• Subject: hypothetical symmetry-bias claim for the comparison side
• Classification: REJECTED
• Reason: No peer-reviewed source locates this specific number. Appears only in marketing pages with no traceable primary paper.
• Action: Remove from layout, or replace with a QUALITATIVE-ONLY descriptor drawn from a related verified study.

Fabrication Risks Detected

List any patterns that suggested numbers were being invented, e.g.:

• Symmetry bias. The right-hand side of the versus comparison has only 2 VERIFIED claims to the left's 5. Planned layout will be asymmetric; no filler statistics will be invented.
• Meta-source attribution. User-suggested citation to Browning et al. (2014) demoted — it is a review, not the primary source for any specific percentage.
• Geographic misattribution. User suggested "Finnish studies" for cortisol research; replaced with correctly attributed Japanese Shinrin-yoku literature.

Recommended Final Claim Set

A cleaned list, in the order they should appear on the poster, ready to hand off to the Layout Spec (Phase 5):

1. claim_id: example_claim_1 — VERIFIED — +15%
2. claim_id: example_claim_2 — QUALITATIVE-ONLY — "Significant reduction"
3. (example_claim_3 dropped)

User Checkpoint

Approval required before proceeding to layout and rendering.

User options:

• Approve this claim set as-is
• Request additional sources for: <claim_id>
• Reshape poster: <describe>
• Accept qualitative language in place of missing percentages