Fact-Checking Policy
Every story on this site is checked before it publishes. This page describes how, in enough detail that you can judge whether we do.
Last updated: August 2026.
Who checks a story before you see it
The reporter checks their own work first, and that is not a formality. Before filing they go back through the piece claim by claim and mark where each one came from: a document, a dataset, a named source, a recording. Anything they cannot source is cut before an editor sees it.
Then a second person reads it. Every story is read by an editor who did not write it, and the rota keeps that true at odd hours. That editor’s job is to be unconvinced. They ask where a number came from, whether the person quoted was in a position to know, and whether the headline claims more than the copy supports, which is where our near misses happen.
Some stories get a third read: anything alleging wrongdoing, anything about a named private individual, anything carrying an evacuation instruction, and anything resting mainly on anonymous sources. There the checking editor works from the underlying material rather than the reporter’s account, and the story waits. We have held pieces overnight and lost them by morning. That is what the rule costs.
Primary sources, and reading the original
We work from the thing itself: the budget document, the court ruling, the agency statement, the person who was in the room. A secondhand report is a lead and not a source. Where we cannot reach the original, the story says what we relied on instead.
Indonesian-language material is read in Indonesian. English aggregation of Indonesian officialdom fails in specific ways. Administrative terms flatten, so a district becomes a province and a recommendation becomes an order, and figures in the local convention arrive an order of magnitude out. When we cite the meteorology and geophysics agency or the disaster management agency, we name it and give the time the statement was issued.
Links, so you can check us
Where a claim rests on a public document, we link it in the sentence making the claim, not in a pile at the bottom and not to a homepage that happens to contain it. Where the source is a long PDF we say which section. If an official site later removes a page, we say when we read it.
Some sources cannot be linked: a leaked file that would identify who gave it to us, or an interview given on condition of anonymity. There we describe the material as precisely as safety allows and say why there is no link, which is more use than “documents seen by us” and nothing more.
Numbers, studies and statistics we are handed
Statistics are checked against the original release, not the press summary. We look at what was measured, over what period, and whether the comparison is with the last quarter or the same quarter a year earlier, because that choice is often the whole story. When the national statistics agency changes a method, we say the series changed rather than reporting a jump.
With studies we read the paper where we can get it, check the sample size and who paid for it, and ask whether the finding supports the claim in the release. A company’s survey of its own users is reported as exactly that, and a funding round is attributed to the startup announcing it unless a filing confirms the amount. Rupiah figures carry a stated rate and date, and we never assert a total we cannot show the working for.
Photographs, video and things that arrive as screenshots
Images we did not take are checked before they run. We reverse image search them, read the file metadata where it survives, and try to reach whoever took it. We look for what gives an old photograph away: signage, plate formats, and whether the shadows agree with the time claimed. Disaster footage recirculates for years here, and a clip from a 2018 tsunami is back within hours of the next earthquake.
Social posts get the same treatment, plus the question of whether the account is who it appears to be. We contact the person before quoting them, ask permission where a post was not clearly public, and say in the copy when a claim is theirs and unverified. Fake ministry circulars travel widely as WhatsApp images, so a screenshot of a document is not a document. We confirm with the issuing body that the text exists, and if we cannot, we leave it out.
The same standard whoever is speaking
A claim is checked the same way regardless of who makes it. A minister, a governor, a company, an activist group and an account with a large following all meet the same question: show us. We do not soften a check because a source is useful to us, or sharpen one because the subject is unpopular this month.
Nothing enters a headline as fact because a powerful person asserted it. A statement is reported as a statement with the speaker named, and where we can establish it is untrue we say so in our own words rather than parking two claims side by side and calling it balance. Where the evidence is unsettled, we describe the disagreement.
When a check fails after publication
Sometimes this process misses something. The story is corrected in place with a dated note, as set out in our corrections policy, and the miss is logged internally with its cause. That log is why we read original Indonesian statements instead of English summaries, a rule that came out of an evacuation story we got wrong.
If you think something we published does not stand up, write to support@wisatahits.blog with the link and the sentence. Editors read and sort that inbox themselves. We would rather argue with a reader about a source than leave a wrong figure on the site.
The standards we work to
We work to the principles the International Fact-Checking Network publishes for its signatories: non-partisanship, transparency about sources, funding and method, and an open corrections policy. We are not a signatory and do not claim to be.
On funding, our ownership and funding page is that disclosure, and our editorial standards cover sourcing and independence. These rules were written by the people who follow them, and get revised whenever we find a hole. The date at the top moves when they do.
