Orientation, impartiality and trust
Endorsements, regulators’ rulings and trust surveys: three kinds of evidence about the news, kept separate.
Why there is no bias rating on this site
Researchers do measure slant — by coding articles, comparing the language outlets use with politicians’ language, or surveying how audiences place outlets — and some of those methods are reproducible. Each measures something specific and depends on choices about what counts as neutral, so their results differ.
This site does not publish a single bias rating. It uses three kinds of evidence that can each be checked against a dated primary record: what an outlet endorsed, what the regulator ruled, and what audiences say in surveys. Where a content-analysis study is used, as for the 2016 referendum coverage, its sample and method are stated.
Editorial orientation
What an outlet said it wanted, in its own leader column, on a dated occasion. An endorsement is a fact.
Regulatory findings
What Ofcom decided, in a published adjudication, about a broadcaster’s due impartiality. A breach is a fact.
Audience trust
What surveyed readers and viewers say. A fact about opinion, reported as such.
None of the three is “bias”. Together they are what can honestly be said.
What the national papers actually backed
Two documented decisions per title: the position taken at the 2016 referendum, as classified by King’s College London’s content analysis, and the endorsement given at the 2024 general election. Front-page counts are referendum lead stories over the 10-week campaign.
The Daily Telegraph / Sunday Telegraph
2016 referendum: Leave2024 election: ConservativeReferendum front-page leads: 41Published more referendum front-page leads than any other title in the KCL study.
Daily Express / Sunday Express
2016 referendum: Leave2024 election: ConservativeReferendum front-page leads: 33Daily Mail
2016 referendum: Leave2024 election: ConservativeReferendum front-page leads: 25The Daily Mail backed Leave while its Sunday sister title backed Remain — under the same owner, the same week.
The Mail on Sunday
2016 referendum: Remain2024 election: ConservativeThe Sun / The Sun on Sunday
2016 referendum: Leave2024 election: LabourReferendum front-page leads: 22Switched to Labour on 3 July 2024, its first Labour endorsement since 2005.
The Sunday Times
2016 referendum: Leave2024 election: LabourBacked Leave in 2016 and Labour in 2024 — the same publisher as The Times, which did neither.
The Times
2016 referendum: Remain2024 election: No endorsementReferendum front-page leads: 18The Guardian / The Observer
2016 referendum: Remain2024 election: LabourReferendum front-page leads: 24The only Remain-backing title to publish more than four immigration front-page leads during the campaign.
Financial Times
2016 referendum: Remain2024 election: LabourReferendum front-page leads: 16Its 2024 endorsement was its first for Labour since 2005.
Daily Mirror / Sunday Mirror
2016 referendum: Remain2024 election: LabourReferendum front-page leads: 5Published five referendum front pages to the Sun’s 22 — the starkest asymmetry in the study.
The i Paper
2016 referendum: No position2024 election: No endorsementHas a standing policy of never endorsing a party. Now owned by DMGT, which also publishes the Daily Mail.
2016 positions and front-page counts: Moore, M. and Ramsay, G., UK media coverage of the 2016 EU Referendum campaign ↗. The study excluded the Independent, the Daily Star and the Sunday People — see its own inclusion rules. 2024 endorsements: each paper’s own leader column, compiled by Press Gazette ↗.
Reading the endorsement table
Three things in that table that break the standard story
The Daily Mail backed Leave while the Mail on Sunday backed Remain — same owner, same week. Reach plc owns both the Mirror, which backed Remain, and the Express, which backed Leave. And in 2024 most of Fleet Street, including the Sun, endorsed Labour. Positions move; ownership mostly does not. Any theory in which proprietors dictate a house line has to account for all three.
Newspapers may campaign. Broadcasters may not.
This is a legal distinction, not an editorial judgement, and it is the single most-ignored fact in arguments about UK media bias. Everything in the endorsement table above is entirely lawful.
Newspapers
No impartiality duty
No legal impartiality obligation of any kind. A newspaper may endorse a party, campaign for an outcome, and select which stories to run on that basis. Accuracy codes apply; political neutrality does not.
Regulator: IPSO or Impress, both voluntary; most nationals belong to IPSO. The Guardian and FT belong to neither and run their own systems.
Broadcasters
Impartiality required
Bound by Section Five of the Broadcasting Code. News must be reported with due accuracy and presented with due impartiality, and due impartiality must be preserved on matters of political or industrial controversy.
Regulator: Ofcom, statutory
Platforms and search
No impartiality duty
Not covered by the media ownership rules and not subject to any impartiality duty, despite being used for news by six in ten UK adults. The Online Safety Act gave Ofcom powers over harm, not over the plurality of what people see.
Regulator: Ofcom, but only for online safety duties — not for plurality or impartiality
The misconception this settles
“Due impartiality does not mean an equal division of time has to be given to every view, or that every argument and every facet of every argument has to be represented.”
Ofcom Broadcasting Code, Section Five — read the section ↗
“Due” means adequate or appropriate to the subject and the nature of the programme. A documentary on the safety of a licensed medicine does not have to give equal time to the view that it is poison, and a news bulletin does not have to balance every government statement with an opposition one within the same item. Complaints that a broadcaster gave "more airtime to one side" therefore very often fail — not because the regulator is captured, but because equal airtime was never the test.
The rules that actually apply
Rule 5.1
News, in whatever form, must be reported with due accuracy and presented with due impartiality.
Rule 5.5
Due impartiality on matters of political or industrial controversy and matters relating to current public policy must be preserved — this may be achieved within a programme or over a series of programmes taken as a whole.
Rule 5.7
Views and facts must not be misrepresented. Views must also be presented with due weight over appropriate timeframes.
Rule 5.9
Presenters may express their own views on matters of political or industrial controversy, provided alternative viewpoints are adequately represented.
Rule 5.11
In dealing with major matters of political or industrial controversy, an appropriately wide range of significant views must be included and given due weight in each programme or in clearly linked and timely programmes.
What the regulatory record shows
Ofcom has recorded due impartiality breaches against broadcasters across the political spectrum, including both the BBC and GB News. Any account of broadcast impartiality that finds fault in only one direction is not describing the regulatory record.
And ignore the complaint totals. Complaint volumes are not a measure of bias. They are not adjusted for hours broadcast, and one organised campaign can add thousands of complaints about a single programme without changing whether a rule was broken. What carries information is the adjudication.
Read Ofcom’s published broadcast standards decisions ↗Survey questions, samples and comparison limits
Ofcom 2026: UK adults aged 16+, November 2025–March 2026; 5,397 interviews (3,718 online-panel, 1,679 face-to-face), effective weighted sample 4,288. D1–D22 ask which named media people saw or heard news on in the last month. Multiple selections overlap. Quotas and weights cover age, gender, geography, social class, urbanity, working status, ethnicity and internet use. Ofcom tests differences at 99%; its approximate tolerance at 50% is ±2 percentage points for the full effective sample, not every subgroup. E3 trust ratings use each platform’s news users: TV 4,077; radio 2,716; print 1,404; social/video 2,927. Recall, panel selection and mixed-mode effects remain. Changed questions, passive-use prompts and continuous fieldwork mean no direct comparison with 2025. Technical report.
Ofcom 2025: UK adults 16+, 4,573 respondents; 5 November–6 December 2024 and 3–28 March 2025. Mixed online and face-to-face quota samples; demographic, nation and mode weights. C1 asks which platforms people use for news “nowadays”. Age-chart bases are 585 aged 16–24 and 472 aged 75+. Ofcom uses effective samples and 99% significance tests; small subgroup differences are not precise rankings. Attention estimates also use industry audience measurement and are not simply survey reach. Technical report.
Reuters Institute 2026: YouGov online non-probability panel, UK adults 18+, sample 2,052; main fieldwork mid-January–end February 2026. Age, gender, region and education quotas; weights use census/accepted population targets. Trust is agreement with “I think you can trust most news most of the time”; it measures opinion. This design has no conventional polling margin of error; differences of two points or less warrant particular caution. Older and less affluent offline populations are under-represented. Since 2021 people reporting no news use in the previous month are included: earlier years excluded them, creating a small composition break. Method and questionnaire.
What audiences actually think
Questions behind these figures
Reuters Q6_2016 records agreement about trusting most news; Q6_2018 records disagreement about trusting social-media news. Q7a asks about paying for, or accessing paid, online news during the last year. The TV comparison concerns reported news use. All use the UK online sample described in the survey methods above, not an audit of news accuracy.
Original questionnaire30%
trust news overall
Down 5 points in a year; the global average is 37%.
73%
distrust news on social media
Attitude towards news on social media, not a count of false stories.
10%
pay for online news
One in ten. The business model problem behind most of this section.
12pp
fall in TV news use since 2021
Change in self-reported TV news use; this is not all television viewing.
Higher-trust brands among those surveyed
- BBC News
- ITV News
- Channel 4 News
- Financial Times
BBC News, ITV News, Channel 4 and the Financial Times are among the more trusted brands included in this survey. The selected list is not exhaustive, small differences should not be ranked, and trust is not a measure of reach or causal influence.
Trust in news overall fell five points in a year, to 30% — below the 37% global average. But the Reuters Institute notes that trust in individual news brands barely moved. This distinguishes general trust from ratings of named brands; the survey does not establish why attitudes changed.
Source: Reuters Institute Digital News Report 2026 — United Kingdom ↗. An online survey of news consumers — independent analysis, not an official statistic.
Coverage is not persuasion
The limit on everything above
The KCL study measured how much was published, how prominently, and how issues were framed. It did not measure whether any of it changed a single vote, and it does not claim to. Anyone citing it as proof that the press delivered Brexit is citing it for a finding it does not contain.
Next: the referendum case study, where the coverage was measured in detail, and who funds the parties.