Methodology
How we source statistics, fact-check claims, and handle updates. This is the editorial standard we hold ourselves to.
One-line version
We aggregate; we don't survey. Every statistic links to its primary source. Pages are dated, and we refresh statistics roundups quarterly.
Sources we trust
Our statistics pages prioritize, in roughly this order:
- Government statistics bureaus — US BLS, BEA, NCES, CDC, NIMH, EPA, NOAA, NASA; UK ONS; Eurostat; OECD; IEA; IPCC.
- Independent research institutions — Stanford HAI, Pew Research Center, Kaiser Family Foundation, Brookings, RAND, Mental Health America, AFSP, Federal Reserve and FRED data series.
- Peer-reviewed publications — Nature, Science, NEJM, Lancet, JAMA, and similar.
- Industry research with disclosed methodology — McKinsey, BCG, Bain, IDC, Gartner, Chainalysis. Cited when methodology is disclosed and the firm has a track record; treated with more caution than government or academic sources.
- Primary corporate filings — 10-K reports, earnings calls, investor presentations. Used for numbers companies disclose themselves.
We avoid: aggregator sites that obscure original sources, anonymous "industry insider" reports, paywalled summaries we can't verify, and any source whose methodology isn't disclosed.
How a statistics page gets written
- Topic selection — based on search demand, citation potential, and gaps in existing 2026-dated coverage. We don't write roundups on topics where Wikipedia, government agencies, or established explainer sites already do a definitive job — only where there's a clear opportunity to do better, fresher, or more accessibly.
- Source collection — we identify 8–15 primary publications, prefer the most recent edition, and pull headline numbers with their methodology context.
- Drafting — every claim is paired with a source link. Numerics include the year and methodology source.
- Editorial review — a human editor verifies each numeric claim against the cited source.
- Publish — with the publication date stamped in the page header and structured data.
- Quarterly refresh — we re-check each statistic against the most recent source release. Outdated numbers are updated and the
dateModifiedfield is bumped.
How an explainer article gets written
- Pick a clear definition that works as the opening sentence — what we'd want an AI search engine or a 14-year-old reader to quote.
- Outline the structure: definition, why it matters, how it works, real-world examples, FAQ.
- Draft. AI tools are used in drafting and research, but every article is reviewed by a human editor.
- Add 2–5 external links to authoritative sources, plus 5–10 internal links to related articles.
- Write the FAQ — at least three questions, each answer self-contained and citation-worthy.
- Editorial review for accuracy, voice, and humanization compliance (we have an explicit banned-words list to avoid corporate-blog clichés).
- Publish, monitor reader feedback, update as needed.
Limitations we want you to know about
- We aggregate. We don't conduct primary research, run surveys, or commission studies. Everything cited is from someone who did.
- Statistics decay. A 2026 stat is current today. Three years from now, the page should either be updated or marked stale. Our quarterly refresh policy is the mitigation.
- Source citation is best-effort. Some industry research is published in PDFs without stable URLs. We link to the most stable source we can find and note when a citation is to a snapshot.
- We are not the original source. If you need to cite a number for academic or legal purposes, cite the primary source we link to — not WhatIs.site.
AI use disclosure
We use AI tools for drafting, structure, fact-finding, and editing — including state-of-the-art large language models like Claude and GPT-4o. Every article and statistics page is reviewed by a human editor before publication. Numbers are verified against named primary sources. We do not publish AI-generated content unchecked.
Corrections
Email hello@whatis.site with the URL and the specific claim you'd like reviewed. Include the source you're using if you have one. We respond within a few business days. Statistics corrections happen out-of-cycle (not waiting for the quarterly refresh); explainer corrections may be batched but are always acknowledged.
How to cite us
Every statistics roundup includes APA and plain-text citation snippets at the bottom of the page. For explainer articles, a generic citation:
WhatIs.site Editorial. (year). Article title. WhatIs.site. https://whatis.site/slug
Standard academic-and-editorial use is welcomed without permission. For bulk republication or derivative datasets, please email first.