# What is media bias? Seven types, and the ones readers miss

> Most people check news for a left or right lean, and that's one of seven kinds of media bias. The seven types, what the studies found about each, and six checks that work on any story.

Source: https://informed.now/blog/what-is-media-bias

By Kira Shishkin · October 7, 2026 · News Curation

**Summary:** Media bias is any steady tilt in what news covers, how it frames it, and whom it quotes, one that pulls the picture a reader gets away from the events themselves. Most people check for a partisan lean, yet a review of 59 studies of US election coverage found almost none in newspapers. The types every outlet shares come from how news gets made: Selection, framing, and false balance. Each one has a check you can run on any story.

Ask whether an outlet is biased and most people answer with a direction, left or right. That's the bias partisans argue about, and it's the narrowest kind. Long before a reporter picks an adjective, someone has decided which events count as news, which story leads, which expert gets a call, and what the problem is called. Any of those choices can change what readers believe, and none of them needs a political motive.

### What media bias means

Media bias is a consistent tilt in how news gets selected or presented. Sometimes the tilt is political. More often it's a habit of the trade, like favoring conflict or quoting two sides, that runs the same way no matter who's in power.

Most Americans believe they see it. In a September 2024 Pew Research Center [**survey**](https://www.pewresearch.org/short-reads/2024/12/02/most-americans-continue-to-say-media-scrutiny-keeps-politicians-from-doing-things-they-shouldnt/) of 9,680 US adults, 77% said news organizations tend to favor one side on political and social issues, close to the highest share since Pew first asked in 1985. Republicans said so at 88% and Democrats at 67%, and 22% said outlets deal fairly with all sides. The question asks about sides, which matches how most people picture bias. The research on how news changes minds points to seven types, and most of them work the same way whichever side an outlet is on.

### Seven types of media bias

**Selection.** The first choice is which events become stories at all. In 1949, journalism professor David Manning White asked the wire editor of a morning paper in a Midwestern city of 100,000, whom he called Mr. Gates, to keep every wire item he rejected for a week and write down why. White's [**study**](https://doi.org/10.1177/107769905002700403) found Gates used about one item in ten. Roughly two-thirds of the rejections came down to space, and the rest were his own calls about what was worth printing, a few of them political. Every newsroom still runs that filter, now guided by about seven [**criteria**](https://informed.now/blog/how-news-gets-chosen) that favor conflict, prominence, and the unusual, and an algorithmic [**feed**](https://informed.now/blog/curated-news-vs-algorithmic-feeds) runs it a second time with a model.

**Framing.** How a story defines the problem, its cause, and the fix. Communication scholar Robert Entman's 1993 [**definition**](https://doi.org/10.1111/j.1460-2466.1993.tb01304.x) is the standard one: To frame is to "select some aspects of a perceived reality and make them more salient." In 1997, political scientists Thomas Nelson, Rosalee Clawson, and Zoe Oxley [**tested**](https://doi.org/10.2307/2952075) what that does to an audience. They had people watch one of two local news stories about the same Ku Klux Klan rally, one framed around free speech and the other around public order. Viewers of the free speech version reported more tolerance for the Klan's rally than viewers of the public order version, and a second experiment on a simulated electronic news service found the same result. The frame worked by changing how much weight viewers gave to public order.

**Placement.** Which story leads, and which fact lands in the first paragraph or the fourteenth. Entman counts placement and repetition among the main ways a text makes some facts stand out over others, and an editor can use both without changing a single fact. A detail in the last paragraph is technically reported and practically hidden.

**Source selection.** Who gets quoted. A story has room for a few voices, and the choice of which official, expert, or bystander speaks decides whose account sounds like the default. A story built only on official sources tells the official version, accurately.

**Word choice.** The type most people mean when they say bias: "Tax relief" or "tax cuts," "riot" or "protest," and "slams" or "criticizes." It's the easiest type to spot and the one partisan readers notice first.

**Omission.** What's missing from a story that did run: Last year's number next to this year's, the forecast that missed, and the denominator. A crime story that reports this year's count without last year's leaves readers to supply the trend themselves.

**False balance.** Giving two positions equal weight when the evidence behind them is unequal. Maxwell and Jules Boykoff [**sampled**](https://doi.org/10.1016/j.gloenvcha.2003.10.001) 636 global warming articles published between 1988 and 2002 in the New York Times, the Washington Post, the Los Angeles Times, and the Wall Street Journal. In 52.7% of them, the view that humans were driving warming and the view that natural variation explained it got roughly equal attention, during years when scientific assessments increasingly pointed to human causes. A rule meant to keep reporters fair produced a bias of its own.

### Partisan slant is smaller than readers think

In 2000, communication researchers Dave D'Alessio and Mike Allen pooled 59 quantitative studies of US presidential campaign coverage going back to 1948. They sorted partisan bias into three kinds: Which stories ran, how much coverage each side got, and whether that coverage favored one side. Their [**meta-analysis**](https://doi.org/10.1111/j.1460-2466.2000.tb02866.x) found no significant partisan bias in newspapers, close to none in newsmagazines, and small tilts in network television news that the authors judged probably insubstantial.

That result sits beside the hostile media effect, the finding that partisans on opposite sides see the same coverage as slanted against them, which our post on [**unbiased sources**](https://informed.now/blog/unbiased-news-sources) covers. Readers perceive more partisan bias than researchers measured, at least in mainstream election coverage. The structural types run through every outlet, left, right, and center, because they come from the craft itself: Limited space, a taste for conflict, and the habit of giving each side a quote.

### Isn't some news plainly partisan?

Yes. The 2000 review covered newspapers, newsmagazines, and network television during presidential campaigns, so it says nothing about cable opinion shows or social feeds. Later work found real slant in print too: A 2010 [**study**](https://www.nber.org/papers/w12707) by economists Matthew Gentzkow and Jesse Shapiro traced the political language of 433 daily newspapers to the politics of their readers. Opinion hosts and partisan sites take sides by design. If most of your news comes from them, partisan slant is your largest bias, and the charts that rate outlets from left to right are worth using.

What we can't settle is which type moves belief most. We don't know of a study that ranks them against each other. We'd still argue that a lean you can see is a lean you can discount, and readers already adjust for a host's politics. Selection, framing, and false balance work on readers who think they've already corrected for bias, and they work the same way in the outlets those readers trust.

### How to spot media bias in a story

These take a few minutes and work on a newspaper, a broadcast, or a feed.

1. **Check what else happened.** Before reading an outlet's top story, scan a wire service's list of the day's stories or a second outlet's front page. Whatever made one list and missed the other is selection, made visible.
2. **Name the frame.** Say in one sentence what the story treats as the problem and whom it treats as the cause. Then try a different frame on the same facts: A rally as a free speech case or a public safety case, a price jump as corporate greed or a supply shortage.
3. **Restate the headline in flat words.** Swap loaded verbs and labels for plain ones, "says" for "slams" and "rose" for "soared." If the story shrinks after the swap, the words were doing work the facts weren't.
4. **Count the voices.** List everyone quoted, then list who'd have standing to speak and wasn't asked. The second list is where source selection shows.
5. **Weigh each side by its evidence.** When a story gives two positions equal space, ask whether they have equal support. A national science academy and one dissenting researcher make two quotes and very different weights of evidence.
6. **Read to the bottom.** News stories put what editors judged least important last, and that's often the context: The base rate, the earlier estimate, and the caveat. If the last paragraph changes your read of the first, placement was steering you.

### How we handle this at informed.now

Every type on the list applies to us. Our largest is selection: We choose stories by how much an event changes the lives of how many people, a deliberate lean toward consequence and away from spectacle, and our [**methodology**](https://informed.now/methodology) page says so in writing. Every brief [**informed.now**](https://informed.now/) sends names the strongest source behind it, so you can run the voices check and the evidence check on us in a minute.

### The shape of the news

Asking whether an outlet leans left or right is a fair first question and a poor last one. Every story reaches you after choices about what counts, what leads, who speaks, and what the problem is called, and most of those choices would look the same under any party. Bias, in that sense, is the shape of the news. Readers who can see the shape are harder to steer than readers who only check the direction.