What is the most unbiased news source? The two organisations that rate this do not measure the same thing

There is no ranking that answers this, and the reason is more useful than a list would have been.

Two organisations dominate media bias rating in the United States, and they are measuring different properties.

**AllSides** rates political lean and nothing else. Its method combines several inputs: blind bias surveys, in which ordinary readers rate content without being told which outlet produced it; multi-partisan editorial review by trained panels; independent reviewer analysis; and occasionally third-party academic data. Its reviewers look for as many as sixteen distinct types of bias — spin, slant, sensationalism, bias by omission, analysis presented as fact. Outlets receive a category and a numeric score from −6 to +6, where 0.0 is exact centre.

**Ad Fontes** uses two axes. Political bias runs left to right; reliability runs high to low. Trained analysts drawn from across the political spectrum rate individual articles and programmes, and those article-level scores are aggregated into a source position through a weighted algorithm.

That structural difference is the whole answer to the question in the headline.

AllSides states explicitly that **it does not rate accuracy or credibility**, on the grounds that a publication can be entirely accurate and still be biased. So an AllSides centre rating is a statement that content does not lean, and specifically not a statement that it is true. Ad Fontes separates the two deliberately, which is why its chart is two-dimensional: an outlet can sit dead centre on bias and low on reliability, and that combination is not unusual.

Which means "most unbiased" and "most reliable" are different questions with different answers, and the phrasing people search for collapses them.

There is a further limitation worth stating plainly. These ratings are produced by human judgement — panels, surveys, trained analysts — and both organisations say so. They are structured, documented judgement rather than measurement in the way a temperature is measured. Two well-run rating systems applying different criteria to the same outlet can reach different positions without either being wrong, and they do.

One more structural point explains why no ranking settles this. Both organisations rate **outlets**, and readers encounter **articles**. A newsroom rated at the centre publishes individual pieces across a spread, and a strongly-leaning outlet still runs straight wire copy. An outlet-level score is an average over output the reader is not currently looking at, which makes it a poor guide to the specific piece in front of them.

That is also why the two systems can disagree without either failing. AllSides weights reader perception through blind surveys; Ad Fontes weights trained analyst scoring of individual content. Those inputs answer subtly different questions — how does this read to people who do not know the source, versus how does this score against a rubric — and there is no reason they should converge.

So the practical answer is not an outlet name. It is a change of question.

Instead of asking which source is unbiased, the checkable questions are: does this piece say where each claim came from, and can I reach the source it names? Does it distinguish what is established from what is asserted? Does it state what it does not know? Does it publish corrections, and can you find them?

Those are properties of an individual article, verifiable in a few minutes, and they do not require trusting anyone's rating of the outlet — including the ratings described above.