Two reporters can both file accurate pieces about the same event and walk away with two different stories. That gap is framing. The selection of which fact comes first, which person gets named, which detail gets buried — these are the moves that turn a record into a narrative.
News framing analysis is the discipline of doing that comparison systematically — across outlets, on the same event, holding the underlying facts constant so the editorial choices become visible.
Here is how to do it, in five steps, using real coverage from the week of May 15, 2026.
Step 1: Establish the underlying event independent of any single source
Before reading any headline closely, pin down what actually happened — the names, dates, numbers, locations, and direct quotes that are not in dispute. This is the baseline. Everything else is interpretation.
For the Trump-Xi summit that concluded May 15, the undisputed record across the 27 outlets that covered it includes: a two-day meeting at Zhongnanhai, a Chinese commitment to order 200 Boeing jets, Xi's pledge not to supply Iran with military equipment, a Treasury statement that the two countries will "begin AI safety discussions" with no date attached, and Xi's invocation of the "Thucydides Trap." Those facts are the floor.
Everything an outlet adds, omits, sequences, or characterizes beyond that floor is a frame.
Step 2: Compare the lead — what comes first in the headline and the first paragraph
The lede is the single most powerful framing move an outlet makes. Two outlets can put the same fact in the same article and tell two different stories simply by changing which fact sits at the top.
Ukraine-Russia coverage this week is the textbook case. The Guardian's headline on the Kyiv strike and Ukrainian drone retaliation led with the retaliation: "Ukraine attacks Russia with drones after suffering three days of massive strikes." Reuters led with the civilian deaths: "Zelenskiy condemns Russia after strike on Kyiv apartment block kills 24." NY Post matched: "Death toll in attack on Kyiv apartment building now stands at 24."
Same event. Same facts. Three outlets. One read first as Ukrainian military capability, two read first as a Russian war crime. Neither framing is wrong. The choice of which half comes first is the framing.
A second example from the same story: only the NY Post named in its opening paragraph that three of the 24 killed were children. Zelenskyy's public statement included that figure. Every other outlet had it available; only one elevated it to the lead. The most humanizing fact in the coverage appeared as a top-of-piece detail in exactly one outlet.
Step 3: Track what gets emphasized — and what gets buried
After the lead comes the order of the rest. What does the outlet treat as the second-most-important fact? What ends up in paragraph nine? What is mentioned only as a qualifier?
The Trump-Xi summit produced a clean example. Treasury Secretary Scott Bessent announced the two countries would "begin AI safety discussions." The NYT business desk added the specific caveat that Bessent "did not say when these talks would happen." Most other outlets listed the AI commitment alongside the Boeing order as a summit outcome, without the no-date qualifier.
An announcement of future talks, with no timeline, is a statement of intent, not an agreement. The qualifier is the news. Whether an outlet kept it or dropped it determined whether the reader walked away thinking a deal got done.
This is also where word choice matters. The same week, in coverage of the Powell-to-Warsh Federal Reserve transition, the New York Times used the phrase "regime change" in its headline — not a personnel swap, but an institutional rupture. That single phrase asks a different question of the reader than "new Fed chair takes over" does. The facts are identical; the frame is the question.
Step 4: Note what is missing across the political spectrum
Framing analysis is not only about how stories are told. It is also about which stories are told to which audiences. A pattern of absences is itself a frame.
Three examples from this week's coverage.
The Fed transition was, for years, a stated priority of right-leaning outlets that spent 2022 through 2025 attacking Powell. The transition they advocated for is now happening — and right-leaning outlets are largely absent from covering it. Fox Business filed one piece, on a Fed governor's resignation that the other five outlets in the story missed entirely.
A global bond market selloff drove US borrowing costs to multi-year highs on May 15. Bloomberg, Reuters, and the Financial Times covered it. No right-leaning outlet touched it. A story about soaring borrowing costs is exactly the kind of fiscal-discipline narrative right-leaning financial press typically runs hard.
The CDC's monitoring operation for 41 people exposed to hantavirus after three deaths on a cruise ship was covered by The Hill and the NY Post. Left-leaning outlets with dedicated health beats did not run it.
Selection isn't an accident. It is a frame. We made the broader case in our piece on why source count matters more than any single article, and the diagnostic side — how to recognize a blindspot when you're inside one — is in our explainer on what a news blindspot is and how to find yours.
Step 5: Look for language that smuggles in interpretation
Once you have the lead, the emphasis, and the omissions, the last layer is the words themselves. Watch for phrasings that import a conclusion the underlying facts have not earned.
The Trump-Xi summit produced two side-by-side examples. Trump claimed "fantastic trade deals." The only countable commercial outcome anyone could name was the Boeing jet order. The NYT's business desk treated the gap between the claim and the record as the news. Most outlets quoted "fantastic trade deals" as a description of the summit's outcome — turning Trump's gloss into a headline fact. That is interpretation traveling as quotation.
The same summit produced the Xi warning about the Thucydides Trap — a head of state telling another head of state on camera that their countries risk armed conflict. Time and the Guardian treated it as a geopolitical signal requiring explanation. Fox News and The Hill repackaged it as Trump's attack on Biden's record: "Trump says Xi agreed US became a declining nation during Biden years." Xi's actual warning about war risk was the news. The Biden attribution was Trump's gloss on it. The choice to lead with the gloss is the frame.
Word-level moves like these are not always partisan. AP's headline on the Kyiv strike read "killing 9" while every other outlet had the confirmed toll of 24. The likely cause is a publication-timing artifact — AP filed early, the headline did not get updated. But a reader who saw only that headline walked away with a number less than half the actual count. Mechanical errors and editorial choices both end up framing the story the same way: by what shows up in the words a reader actually sees.
Putting it together
The five steps in practice:
1. Pin down what is undisputed. 2. Compare the lead across outlets. 3. Track what each outlet emphasizes versus buries. 4. Note who didn't cover it. 5. Watch for words that carry a conclusion.
A reader running this on a single story from a single outlet learns very little, because there is nothing to compare. The method only works across outlets — and the work of pulling six versions of the same event onto one screen is exactly what Signal/noise is built to do.
If you want to try the method on a live story, open any story page on Signal/noise and run the five steps yourself. If you want a worked example focused on two specific outlets, our step-by-step guide to comparing CNN and Fox News coverage walks through the same logic on a single pair. And for the underlying mechanics — selection, omission, emphasis — see our explainers on selection bias in news and bias by omission.
Framing analysis is not a way to decide which outlets to trust. It is a way to see what each one is doing — so trust becomes a position you arrive at rather than a default you inherit.