A media bias chart is a satisfying thing to look at. AllSides released version 11 of its chart this spring; Ad Fontes Media now plots 137 outlets on its grid. Each source gets a dot — left, center, or right, sometimes with a reliability score stacked on top. It feels like a map of the whole press in one image.
It is a useful map. It is also a still photograph of a moving thing. A dot tells you where an outlet tends to sit on average, across months of output. It cannot tell you how that outlet framed the story you read this morning. That gap — between an outlet's average lean and the specific article in front of you — is where most of the interesting information lives, and it is exactly what a chart is built to hide.
Here is what a media bias chart leaves out, using real stories from Signal/noise this week.
A chart rates the outlet, not the article
The clearest limit of any bias chart is that it scores an outlet, not the piece you are actually reading. The average is real, but no single article is the average.
Take the Supreme Court's ruling backing the Trump administration on immigration. Within the left-rated outlets covering it, the framing ranged widely. The New Republic's headline called the ruling "gutter racism." The Guardian's lede described "a pitiless campaign against people of colour fleeing violence and disaster" and cast the Court as Trump's "loyalists." A chart would file both outlets under "left" and stop there. It would not tell you that one reached for the most charged language available while the other built its frame into the opening sentence. Those are different editorial choices, and a dot on a grid flattens them into the same color. You can see the full source spread on the Signal/noise story page for the ruling.
Now the reverse. John Bolton pleaded guilty to mishandling classified documents, and the most complete version in our set came from The New York Times — an outlet most charts place on the left. The Times noted that the investigation "spanned the Trump and Biden administrations," the one detail that makes the case politically awkward for everyone. If you had used a chart to decide which outlet to trust on that story, the rating would have steered you wrong about that specific article. The label is an average; the article is a data point. This is the heart of why bias ratings alone aren't enough to read the news critically.
A chart can't see the frame
Two outlets can report the same facts and still tell two different stories. That choice — what kind of story this is — is framing, and it is invisible on a left-right axis.
When Mamdani-backed socialists swept New York's primaries and a rent freeze was approved, Salon's frame was explicitly forward-looking: the institutions behind the surge are already backing candidates around the country. That turns a local primary result into a movement story. A governance-focused outlet covering the identical result tells a story about rents, budgets, and city council math. Same vote count, two different stories — one about momentum, one about administration. The chart says "Salon, left" and moves on. It never registers that the more consequential decision was the kind of story being told. We walk through how to catch this in our five-step method for analyzing news framing.
A chart can't tell you what's missing
The most important thing about a news story is sometimes that one side didn't run it at all. A bias chart, by design, only rates coverage that exists. It has no way to represent absence.
Two stories this week were defined by who skipped them. When Strait of Hormuz shipping rebounded after the U.S.–Iran deal and oil tankers began flowing again, left-leaning outlets published nothing on the reopening — even though many had covered the closure heavily. And when Volkswagen announced plans to cut up to 100,000 jobs and shut plants under pressure from Chinese competition, right-leaning outlets were entirely absent, despite that being the kind of story conservative coverage usually leans into. Neither gap shows up on any chart. Both are exactly the kind of omission we cover in what bias by omission is — and you can see the lopsided source counts yourself on the Hormuz reopening story.
A chart can't see placement
Even when every outlet covers a story, they don't all give it the same weight. Burying a story on a section page is an editorial decision that shapes what readers absorb, and it leaves no trace on a bias rating.
When Iran attacked a cargo ship near Oman and asserted authority over the Strait of Hormuz, the detail that Iran claimed legal authority over the strait was, in our read, buried by four of the outlets covering it. That is a placement and emphasis choice — the fact was technically present but positioned to be missed. A chart measures the direction of coverage, never its prominence.
How does framing actually affect what readers believe?
This is the practical stakes question, and it's why the gap matters. Decades of research on framing — the way a story is packaged, not just its slant — show that emphasis and word choice move how people understand an issue, often more than the underlying facts do. A reader who sees the immigration ruling framed as "gutter racism" and a reader who sees it framed as a procedural win walk away with different mental models, even if both articles report the same decision. A bias chart can confirm both outlets lean the same way. It cannot tell you that their framing pulled readers in opposite emotional directions.
Use the chart as a starting point, not an answer
None of this makes bias charts useless. Knowing an outlet's average lean is a fine first move — it tells you what to watch for. The mistake is treating the dot as the conclusion. The lean is the hypothesis; the article is the evidence; and the only way to close the gap is to read the same story across several sources and compare the actual language, the actual emphasis, and what each one left out.
That comparison is the whole idea behind Signal/noise: instead of a static rating, every story shows the real spread of who covered it, how they framed it, and where the blindspots are. If you want to go deeper on why a single rating was never going to be enough, our piece on why source count matters more than any single article makes the case in full. A chart hands you one number. The story usually has a dozen.