Biased News Examples: 5 Mechanisms You Can Check Yourself

Search results for “biased news examples” usually turn up lists that rate whole outlets as “left” or “right.” This page does something narrower: it walks through five specific, checkable mechanisms that produce a slanted impression in a single article, using illustrative examples, and links each one to the NTN tool that surfaces that pattern with a receipt in real tracked articles.

A note on the examples below: they are deliberately hypothetical — constructed to illustrate a mechanism, not to describe or accuse any specific named publication. We do not assert that any real outlet acted in bad faith. The point of this page is the mechanism itself: something you can verify against a primary source regardless of who published it.

1. Selective quoting (contextomy)

Contextomy means quoting a real fragment of what someone said in a way that changes or narrows its meaning, by leaving out a qualifying clause that sat right next to it.

Illustrative example: a public official says, “Crime is down significantly this year, though we still have work to do in a few neighborhoods.” A hypothetical article’s headline and lead quote only the phrase “we still have work to do,” dropping the first clause.

Every word in the shortened quote was actually said. What changed is the impression a reader forms from it. You can check this yourself by comparing any quoted fragment to the full remarks — a transcript, a press-conference recording, a court filing.

NTN’s version: Quote Check lines up a published quote next to the full primary-source passage it was drawn from and marks whatever text sits between the quoted portions, with a link to the original source.

2. Headline vs. body mismatch

A headline can make a claim that is stronger, more definite, or differently framed than what the article’s own reporting actually supports — and most readers only ever see the headline.

Illustrative example: a headline reads “Company Ordered to Pay Millions,” while the body reports that a court denied a motion to dismiss — a preliminary procedural step, not a final judgment or damages order.

Both the headline and the body can be defensible readings of a fast-moving story. The gap between them is what’s worth checking, and it’s something you can check by reading past the headline and comparing the specific claim to the specific evidence cited for it.

NTN’s version: Headline-vs-Body Check extracts the claim a headline is making and shows the body sentence that does or doesn’t support it, at a stated confidence level.

3. Selection and omission — who covers what

Slant isn’t only about how a story is worded — it’s also about which stories get covered at all, and by which outlets. One outlet covering a story and a comparable outlet not covering it is a different kind of signal than any single sentence.

Illustrative example: a local ballot measure passes. Several outlets in the same market report the result the next morning; one comparable outlet never publishes a piece on it.

One omission proves very little on its own. A pattern of consistent omission across many comparable stories is something you can only see by tracking a fixed set of outlets over time, which is what a coverage ledger is for.

NTN’s version: Coverage-Presence Ledger records, within a defined set of outlets, which ones published a matching story and which did not, with capture timestamps.

4. Loaded word choice

The same underlying fact can be described with words that carry different connotations — “slammed” versus “criticized” versus “responded to,” or “slashed” versus “reduced” versus “cut.”

Illustrative example: a city council votes to reduce a department’s budget by four percent. One hypothetical version says the council “gutted” the budget; another says it “trimmed” it. Both may describe the same four-percent figure.

Word choice like this frequently changes between an article’s first published version and a later edit of the same piece — which is exactly the kind of change a revision tracker is built to catch and display side by side.

NTN’s version: the Revision Tracker captures article text on a fixed schedule and publishes a full diff whenever a captured version changes, so you can see precisely which words were swapped and when.

5. Decontextualized figures

A number can be entirely accurate and still misleading if it’s presented without the comparison point that would let a reader judge whether it’s large, small, typical, or unusual.

Illustrative example: “Applications rose 300% this quarter” is accurate if the prior quarter had four applications and this one had sixteen — and easy to misread if the base number never appears in the piece.

Figures like this often originate inside a quoted official statement or press release before making their way into an article’s own text, which means the same tool that checks quotes for omitted context can often check the figures inside them.

NTN’s version: Quote Check surfaces the full passage a number was drawn from, so you can see whether a comparison point was present in the original source and absent from the published version.

NTN’s stance: the receipt, not the motive

For every mechanism above, the linked tool does one thing: it shows you the receipt — the original text, what it was compared against, when each version was captured, and a link to the primary source. None of the tools label an outlet’s motive, and none use words like “biased” or “misleading” as a verdict.

Cutting a quote for length, writing a punchy headline on a breaking story, or reaching for a strong verb on deadline are all common, often entirely legitimate editorial choices. The same underlying evidence can reflect normal editing, a rushed deadline, or a deliberate choice — and we don’t decide which. We publish what changed and what it was compared against; you draw the conclusion. See Methodology for how captures are confirmed and reviewed before anything is published.