Methodology
What we measured, from whom, and how many. We never say humanity believes anything. Every figure below carries its sample and its limits, and the figures in this preview are placeholders pending panel ratification.
The selection rule
Facts come from an independent named panel; values come from everyone. The panel establishes the factual terrain of a question and nothing else. Panelists are selected for disciplinary coverage of the question, not for balance of opinion, and every seat carries a published funding and affiliation disclosure. No panelist writes the question, sees the results before publication, or reviews the copy of a claim they did not source.
Panel roster
Panel seat 1 — land use and agriculture
Anchor seat, outreach in progress · Disclosure published on ratification.
Panel seat 2 — conservation biology
Anchor seat, outreach in progress · Disclosure published on ratification.
Panel seat 3 — food systems economics
Open · Disclosure published on ratification.
Panel seat 4 — nutrition and public health
Open · Disclosure published on ratification.
Claims and sources in the live question
Should we eat less meat to protect the diversity of life on Earth?
Roughly half of the world's habitable land is farmed, and about 80% of that farmland is used for livestock.
Source: FAO, 2024
Livestock supplies around 17% of global calorie intake and 38% of protein intake.
Source: FAO Food Balance Sheets, 2023
Agricultural expansion is the recorded driver in roughly 90% of tropical deforestation.
Source: FAO Remote Sensing Survey, 2022
Claims verified: 2026-08-04
Sample composition by source
| Source | Responses | Share |
|---|---|---|
| Earthbeat | 636 | 26% |
| a shared link | 412 | 17% |
| The Guardian | 610 | 25% |
| Le Devoir | 465 | 19% |
| Grist | 361 | 15% |
This is a convenience sample drawn through publisher embeds and shared links. It is not representative of any population and is not weighted. Referring source is recorded on every response so contextual bias can be measured rather than hidden.
Excluded claims
- Any figure we could not trace to a primary source with a publication date.
- Projections presented as measurements.
- Single-study findings without replication, where a synthesis exists.
- Any claim whose strongest form would favour one side of the trade-off.
What we store
An anonymous response id, the answer before the trade-off, the answer after it, a self-declared and optional country, a timestamp rounded to the hour, the locale, and the referring source. There is no field for an IP address, precise location, or anything resembling personal data. The widget sets no cookies and makes no third-party requests.