Editorial Standards & Ethics

This is the working rulebook, not a statement of values. It is written down so a reader can hold us to it, and so nobody on the desk has to guess where the line sits at eleven at night. Last updated: August 2026.

Accuracy, and the story that is still moving

Nothing is published here because it is probably true. Every claim traces back to something we have seen or somebody who told us on the record. Where a claim rests on a document, we read the document, and where it rests on a number, we check that number against the body that produced it.

General news punishes this. A story breaks at nine, three other sites have it by five past, and the honest position is that we know one thing and are guessing at four. So we publish the one thing, name who told us, and say what is not yet established. We never quietly promote a claim from reported to confirmed between edits.

The second kind of error we count

Most newsrooms log an error when a fact is wrong. We log a second kind. If the explanation was accurate and the reader still finished the piece unable to say what the term meant, that piece failed, and it gets rewritten and recorded as a miss whatever the checker signed off.

Two rules carry it. The explanation is written by the writer, in their own sentences, inside the piece, and may never explain one unfamiliar term by reaching for a second one. Before publication a second editor reads with the single job of being the person who does not know the subject, and marks the first place the copy assumes something.

Sourcing, attribution and the anonymous source

Named, on-record sourcing is the default and most of our copy holds to it. Attribution says who, when and on what basis: a regulator’s published notice is not the same thing as a person at the regulator describing that notice. When we have been told about a document rather than shown one, the copy says so.

Anonymity is granted where the information matters, cannot be had another way, and the person carries a real cost for their name being on it. Two people sign that off, the editor who commissioned the piece and the editor who owns standards. We do not extend it to somebody attacking a rival from behind a screen, and the Entertainment desk does not run a source close to the star, because that phrase nearly always means a publicist with a preference about the framing.

Gifts, holdings and politics

We pay our own way. Travel, tickets and hospitality come out of our budget, and where a trip is genuinely impossible on our own money we say so in the piece and name who covered it. Review hardware goes back when the piece is filed, or is bought at market value if we keep using it. Gifts beyond something a person could eat at a desk go back with a note.

Nobody covering Auto, Business, Tech or Gaming holds shares in the companies on their beat or takes speaking fees from them. Staff and regular freelances declare holdings, side work and close family interests touching their patch, and a declared conflict moves the story to somebody else rather than running with a footnote. No donations, no party membership, no campaign work, no candidacy while working here. How anybody votes is their own business and we do not ask.

News, analysis, opinion and anything paid for

Four labels meaning four different things. News reports what happened and explains the mechanism under it. Analysis reaches a judgement and shows the reasoning that got there. Opinion argues a position and says so at the top and in the byline position. Sponsored content is paid for by somebody else, carries the word sponsored, is written outside the newsroom, and stays out of the feeds where a reader looks for our reporting.

Commercial relationships do not reach the desk. Advertisers see no copy before publication, get no veto and no advance warning, and reporters are not told who is buying space in their sector. A live campaign has never bought a company a quieter week here. The categories we turn down are on the advertising page, and the revenue is itemised on our ownership and funding page.

Right of reply

If a piece criticises a person or an organisation, we go to them before it publishes and put the specific allegations to them rather than sending a vague request for comment. They get a deadline a real person could meet, counted in working hours and not in the forty minutes before we wanted to file.

Silence is reported as silence. We write that an approach was made and no answer came, and we do not dress a refusal to comment as an admission of anything. A response arriving after publication is added with the time it landed, and where it shows we got something wrong our corrections policy takes over. Nobody sees the draft or approves their quotes as a condition of speaking to us.

Plagiarism and fabrication

Both end the relationship. A staff writer who passes off another person’s sentences as their own, or invents a quote, a source or a detail, does not keep the job, and a freelance who does it is not commissioned again. We then re-read everything else that person filed and correct whatever we find, in public, on each page.

Using other people’s reporting is a different thing from stealing it. Where another outlet broke a story we name them in the first mention and link to them there, in the copy, not in a credit line at the foot where nobody looks. On the reference side, definitions are written from source material and never lifted from other glossaries.

Artificial intelligence, plainly

We use these tools and would rather tell you where. They transcribe interviews, and a human checks the transcript against the audio for anything load-bearing. They produce a first pass at translation between our English and Spanish copy, which an editor then rewrites, because the glossary years taught us that a mistranslated term does more damage than a missing one. They help sort the failed-search log each evening.

They do not write what you read, and they do not write explanations. We have watched a model produce a confident definition that was wrong in a way only somebody who knew the subject would catch, which is exactly the failure this site was built to fix. Every published sentence has a person who wrote or rewrote it and answers for it, and an editor who keeps a suggested phrasing owns that sentence as if they had typed it first.

Whose voices end up in the copy

A story about a mechanism needs the person the mechanism happens to. When a utility announces rolling cuts, its statement is the easy quote and the household now cooking at five in the morning is the story. We push commissions toward people with direct experience rather than the analysts who answer the phone quickest.

We read our own sourcing once a month, counting who was quoted and in what capacity, and that read is usually uncomfortable. There is no quota, partly because this desk is small enough that one would be paperwork pretending to be progress. What we do instead is refuse a piece where every voice sits on the same side of the thing being explained.

Bylines, and who you can hold responsible

Every article carries the name of the person who wrote it. There is no house byline covering for a rewritten press release, and short items built from an agency wire name the agency. The editor who owns standards signs off anything contested and can hold a piece over a section editor’s objection.

Writers are reachable through the desk at support@significadopedia.com, one address the editors read and sort themselves rather than a row of invented department aliases. How a claim is checked before publication is set out in our fact-checking policy. What happens when a check turns out to have failed is in the corrections policy.