How to compare news coverage of the same story across countries

1 — Pick a story that is domestic in more than one country. A comparison only works when the event is genuinely a story in both places. If it leads in one country and appears as a paragraph in another, what you have found is a difference in news values, not in framing — a real finding, but a different one, and it will not survive a headline-by-headline reading. Start with events that have two home countries: a summit, a trade dispute, a border, an extradition, a match.

2 — Fix the facts with a wire report before you read anyone's version. News agencies sell the same copy to outlets across the political spectrum, which forces a plainer register than any of their customers would choose alone. Read the agency account first and write down what actually happened — who did what, when, and how many. Everything a newsroom then adds, removes or promotes becomes visible by subtraction, and you are no longer trying to work out the facts and the framing at the same time.

3 — Compare the pages before you read the articles. Most of the editorial decision has already been taken by the time a word is read: what leads, how much space it gets, which photograph runs with it, and what is not on the page at all. Look at the top screen of each front page and list the top three stories on each without opening any of them. The list you end up with is the comparison; the articles are the evidence for it.

4 — Then compare headlines word by word. Read the headlines side by side and ask three mechanical questions. Who is the subject of the sentence, and who has been moved into the passive? Which verb is doing the work — and would a synonym change the reader's judgement? Are numbers given as totals, rates or comparisons, and against what baseline? Two headlines can be entirely accurate, share every fact, and leave a reader with opposite impressions on the strength of those three choices alone.

5 — Write down what each page did not carry. Omission is the hardest bias to see, because nothing on the page points at it, and it is the one thing a side-by-side view makes easy. Once you have both lists, the interesting entries are the stories that appear on exactly one of them. Ask what a reader of only that page would not know today — that question, rather than any judgement about which outlet is more truthful, is the whole value of the exercise.

Each of these produces a difference between two front pages that has nothing to do with editorial judgement — which is why a comparison can look conclusive and be worthless. They are listed roughly in order of how often they catch people out.

Timezone and edition drift. You are almost never looking at the same moment. A Tokyo front page mid-morning in Europe has had a full news day; a New York one has barely started. Two pages that appear to disagree about what matters may simply be several hours apart, and an outlet that looks slow may have published and moved on. Compare inside the same working day where you can, and where you cannot, note the offset before drawing a conclusion from it.

Wire copy mistaken for consensus. Four outlets carrying the same sentence is not four newsrooms agreeing. It is one agency and four subscribers. Before treating agreement as meaningful, check whether the wording is identical: identical wording is evidence of a shared source, not of shared judgement. The genuine signal is where a subscriber departs from the wire — the sentence it cut, the paragraph it moved up, the headline it wrote itself.

An opinion page compared with a news page. Front pages routinely promote a column above a report, and digital front pages rarely label the difference clearly. Comparing an editorial in one country with a news story in another produces a disagreement that was there by design and tells you nothing about either newsroom. Check what kind of piece you are reading on both sides before you line them up.

Machine translation flattening the register. Translation preserves facts and loses temperature. Irony, honorifics, the difference between an official's formal title and a nickname, the register of a verb — these are the first things to go, and they are exactly what a framing comparison is measuring. When a translated headline reads unusually flat or unusually strident, suspect the translation before the newsroom, and read a second sentence of the same piece before concluding anything.

The photograph doing the arguing. Two outlets can run near-identical headlines over completely different pictures — a crowd or a leader, damage or a flag, a face or a map — and the reader will remember the picture. If you only compare the words, you have compared half the page. Say out loud what each image is asking you to feel; the answer is usually a great deal more direct than the headline above it.

Sampling the same side twice. Two outlets that lean the same way are not a comparison, and it is easy to pick two without noticing, because the best-known international titles from any country tend to cluster. Check the leaning label before you choose, or start from a prepared left-centre-right comparison so the spread is guaranteed rather than hoped for.

The pairing does most of the work. If you want to see political framing, hold the country constant and vary the leaning — Crosswire has prepared left · centre · right comparisons in 36 countries, which removes the risk of accidentally sampling the same side twice. If you want to see how a story changes as it crosses a border, do the opposite: hold the kind of outlet constant and vary the country. Two general-interest dailies from two countries are comparable in a way that a wire service and a tabloid never are, and the general-interest daily is the one kind of outlet that exists almost everywhere.

Where you need a fixed point, use a public-service broadcaster or a news agency: the first is chartered to cover everyone, the second writes to be resold to everyone, and both give you something stable to measure the others against. Crosswire tracks 907 outlets across 186 countries, so in most markets you have a choice of all three.

Deliberately, there is no example story on this page. Any story specific enough to be useful would be out of date within a week, and the whole point of comparing live front pages is that the material is today's. The shape of a good session is this: twenty minutes, one event, one wire report, two or three front pages open at once, and two written lists — what each page led with, and what each page did not carry at all. Almost everything interesting is in the difference between those two lists.

If you are running this with a class rather than on your own, the education page has three timed exercises built on the same method, and the glossary defines the vocabulary — selection, framing, omission, lateral reading — that the steps above assume.