Fact-Checking Policy

WEAR YELLOW FOR SETH is an independent general-news publication covering business, economy, entertainment, health, politics, science, sports, technology, and world news. This page explains how we decide that something is true enough to print, and what we do when we cannot.

The rule underneath everything here is simple to state and harder to practice: we verify claims against primary sources. Not against a screenshot, not against a rival headline, not against what everyone seems to be saying. Against the original record.

What gets checked

Every story passes through a factual review before publication. The things we check most closely are the ones most likely to be wrong or most damaging if they are:

Numbers. Figures, percentages, dollar amounts, casualty counts, poll results. We trace them to the release or dataset they came from.
Names and titles. Spelling, role, and whether the person is who the story says they are.
Dates and sequence. When something happened, and in what order, because timelines are where errors hide.
Quotes. Exact wording and full context, checked against a recording, transcript, or the document the quote came from.
Official actions. Rulings, votes, arrests, filings, appointments, sanctions. We confirm these happened and that we have described them accurately.

Going to the primary source

A primary source is the thing itself. If a court decided something, the primary source is the ruling. If a company reported earnings, it is the filing. If an official said something, it is the statement, the transcript, or the recording. If a study made a claim, it is the paper and its data.

Second-hand summaries, other people’s articles, and social posts are leads, not evidence. They tell us where to look. They do not settle the question. A screenshot in particular proves almost nothing on its own, because it is trivial to fake or strip of context, so we treat one as a prompt to find the real document rather than as the document.

Viral and social media claims

A lot of what looks like breaking news starts as a post that spreads faster than anyone can check it. Our approach to viral claims is deliberate:

We find the original: the first post, the full video, the account that actually said it.
We check whether an image or clip is really from the event it is attached to, or recycled from somewhere else.
We ask who benefits from the claim spreading, which often tells us where to be skeptical.
We do not repeat a viral claim as fact simply because it is everywhere. Volume is not verification.

When we write about a false or unproven viral claim, we make clear that it is unproven or false, and we do not amplify it in a way that spreads it further.

Breaking news under time pressure

During a fast-moving event, the facts arrive in pieces and some early reports turn out to be wrong. We would rather be accurate than first. In practice that means:

We label a story as developing while it is still unfolding.
We attribute clearly, saying who is reporting what, rather than stating unconfirmed details as settled.
We separate what we have confirmed from what is still claimed.
We update the story as the record firms up, and we correct it openly if an early detail was wrong.

A developing label is a promise to the reader that the picture may change, and that we will change with it.

Evidence and corroboration

For contested or high-stakes claims, one source is not enough. We look for a second independent source, a supporting document, or an on-the-record confirmation. The more serious the claim, and the more harm a mistake would do, the higher the bar we set before publishing. This works hand in hand with our sourcing rules, which are set out in our editorial standards.

Expert input

Some stories, especially in science, health, and economics, turn on technical questions our editors are not qualified to settle alone. In those cases we consult people who actually know the field, and we favor independent experts over anyone with a stake in the outcome. We explain a specialist’s relevant background so readers can weigh the view, and we do not dress up one person’s opinion as settled consensus.

Claims we cannot verify

This is the part that matters most. If we cannot confirm a claim, we do not publish it as fact. There is no exception for a claim that is juicy, popular, or convenient. Our choices when something will not stand up are:

Hold the story until we can confirm it.
Report only the part we can stand behind, and say plainly what remains unverified.
Report the claim as a claim, clearly attributed, when the fact that it is being made is itself newsworthy.
Drop it.

What we will not do is launder an unverified claim into a fact by writing around the uncertainty.

Manipulated and AI-generated media

Fabricated photos, edited audio, doctored video, and synthetic images produced by AI are now part of the landscape. We check the provenance of visual material before we use it, look for signs of manipulation, and try to reach the original. When we report on a piece of manipulated or AI-generated media, we say clearly that it is manipulated or synthetic. We do not present altered material as a genuine record of events.

If you spot a problem

Readers catch things we miss, and we want to hear about it. If you believe we have published an unverified or incorrect claim, tell us through our contact page or write to support@wearyellowforseth.com. Confirmed errors are fixed and logged under our corrections policy. Checking facts is not a box we tick once. It is the job.

Last updated: August 2026