4.2 - Media influence, bias and opinion polls

How election coverage can influence politics

Media influence can work through several mechanisms rather than a single direct instruction to voters. Agenda setting makes some issues more prominent. Framing presents an issue through a particular interpretation, such as treating a spending pledge mainly as investment or mainly as a tax burden. Priming can make an issue a standard by which people judge a party. Persuasion changes a view, while reinforcement strengthens one already held.

Suppose two reports accurately describe a party's NHS pledge. One emphasises shorter waiting lists, while the other emphasises the risk of unfunded costs. Their selection and framing can influence whether a reader assesses the party by its intentions or its credibility. This is an illustrative mechanism, not evidence that either report changed an actual election. Parties use manifestos, broadcasts, interviews, advertisements and social media to seek such effects. A manifesto presents promised action; media decide which promises to amplify and examine.

Coverage of leaders can shape perceived governing competency, the ability to manage public affairs effectively. Interviews and debates may test knowledge and judgement, while short clips can reward confident presentation or magnify an error. However, prior beliefs influence reception: a supporter may excuse conduct an opponent condemns. National results also depend on party attachment, policies, social circumstances, turnout and constituency competition.

The 1997 and 2024 cases illustrate why communication cannot be studied in isolation. New Labour presented credible change after a long Conservative tenure; in 2024 competing opposition parties gained from dissatisfaction with the incumbent. Campaigns could articulate those preferences, but the existence of a campaign message does not prove that the message created them.

If you sit in 2027: Issue 3 combines media, persuasion, bias and polls in one dense bullet and refers to key general elections. Issue 4 separates the media/bias and opinion-poll demands from June 2028. Persuasion and election examples remain useful mechanisms here, alongside influence between elections; shared assessments use that common teaching.

Bias, selection and different media rules

Media bias is systematic favouring of a position through selection, presentation or commentary. It can be explicit, as in an editorial endorsement, or less obvious, through whose voice is heard, which topics lead the news and which facts receive emphasis. Bias should be demonstrated from evidence: disagreement with a story's conclusion does not by itself establish distortion.

Ownership and editorial priorities can shape an outlet's choices. Commercial incentives can favour conflict, novelty or a recognisable personality. Journalistic routines can favour accessible official sources. Audiences also choose outlets that fit their views, meaning that the direction of influence can run from preferences to media consumption as well as from media to preferences.

The rules differ by medium. Under Ofcom's Broadcasting Code, relevant broadcast news requires due accuracy and due impartiality. “Due” means appropriate to the subject and programme; it does not require exactly equal time for every possible position. Election coverage has additional requirements concerning parties and candidates. Newspapers can make partisan editorial choices, and ordinary social posts do not inherit the broadcasting code merely because they discuss politics. Ofcom Section 5 and Section 6, checked 14 October 2026.

Regulation limits some forms of favouring, but does not remove editorial judgement about importance or framing. A defensible argument can therefore distinguish overt partisan endorsement from a disputed selection effect within regulated coverage. The effectiveness of safeguards needs evidence about actual reporting and enforcement, rather than an assumption that a rule guarantees perfect neutrality.

Accountability and influence between elections

Media influence continues after polling day. Investigations can disclose misconduct or failed delivery, giving Parliament, voters and regulators information they would otherwise lack. Sustained reporting can increase pressure for an inquiry, a ministerial response or policy change. This watchdog role supports accountability, although proving that coverage caused a particular decision requires checking the sequence and other pressures.

Media can also give campaigners a route onto the public agenda. An organisation highlighting housing conditions can seek attention between elections, prompting politicians to respond before the next manifesto. Conversely, repeated attention to a scandal can encourage disillusion, a loss of confidence in politics, or mobilise citizens who want reform. Apathy, lack of political interest or motivation, is a different response. Critical reporting does not necessarily reduce participation.

The Leveson Inquiry's report, published on 29 November 2012, examined the culture, practices and ethics of the press. Its existence demonstrates that media themselves can become objects of political scrutiny between elections. The official report record supports that example; it does not measure an electoral effect. Press freedom and the prevention of harmful practices can pull policy arguments in different directions.

Digital communication can bypass an editor: a party or citizen can publish directly, and sharing can carry material beyond its original audience. Recommendation systems, targeted advertisements and self-selected networks can make different users encounter different political material. Rapid distribution also permits correction and challenge, while making misleading claims harder to contain. These are mechanisms with different possible outcomes, not proof that every user lives in a closed information bubble.

The BES distinguishes probability surveys from its internet panel and warns that social media users are not automatically representative of the public. A loud online response therefore cannot be assumed to be a national majority. BES methodological presentation, slides 14–19. If you sit in 2027: the glossary's phrase “no confident” is corrected to “no confidence” in Issue 4. The accurate meaning of disillusion applies in either series.

What opinion polls measure and why they matter

An opinion poll estimates views or intentions in a defined population from a sample. A voting-intention poll asks what people say they would do; a post-election survey asks about reported behaviour; neither is the count of actual votes. An exit poll samples people after leaving polling stations and models likely seats. A demographic profile weighted to the final result describes groups; it is not a fresh prediction of that already known result.

Polls help parties assess issues, test messages and allocate attention. Media use them to describe competition, and voters may use them when considering tactical choices. A perceived lead could encourage supporters, discourage them because victory seems assured, or mobilise opponents. These bandwagon and underdog possibilities point in different directions; the publication of a poll does not establish which occurred. Polls can also create a focus on who is ahead at the expense of policy scrutiny.

Interpret a poll by asking whom it sampled, when fieldwork occurred, how the question was worded, how undecided people were treated and whether turnout was modelled. Weighting adjusts the influence of sampled groups, but cannot automatically repair an absent group or an inaccurate response. Non-response, unrepresentative recruitment, recall errors and late changes can all matter. Smaller subgroups carry more uncertainty than the headline sample.

An illustrative poll putting Party A at 35% and Party B at 33% is not a guarantee that A leads by exactly two points in the population. Sampling uncertainty and other errors matter; a margin quoted for an ideal random sample does not describe every error in a weighted online or quota poll. Nor does a national vote lead map directly onto a seat lead under first past the post. Repeated independent polls and stable trends are usually more informative than one small movement. British Polling Council explanation.

The 2015 pre-election polls underestimated Conservative and overestimated Labour support, prompting an inquiry and changes to transparency about weighting. That verified failure supports scrutiny of methodology, rather than the claim that all polls are worthless or deliberately manipulated. BPC response to the inquiry, 31 March 2016.

Building a supported argument about influence

A correlation between media consumption and party choice is a starting point for explanation, not proof of persuasion. In MORI's 1997 aggregate, 52% of Sun readers and 67% of Guardian readers supported Labour; Conservative support was 30% and 8% respectively. Those national estimates show associations. Readers might already prefer an outlet's position, share characteristics associated with its audience, or encounter many other sources. MORI's 1997 readership table.

To test influence more convincingly, a repeated study could measure preferences before and after exposure, compare similar people with different exposure, and consider rival explanations. Experiments can help isolate an effect, although an artificial setting may not reproduce a whole election campaign. Content analysis shows what was communicated; audience reach shows who could encounter it; neither alone establishes that opinions changed.

Consider this complete original pair of arguments. One argument says repeated critical coverage can change the criteria voters use to judge government, especially where party ties are weak. The other says selective audiences and prior attachments often make media reinforce views, while an incumbent's actual record explains both criticism and falling support. The first identifies a plausible mechanism; the second identifies a rival causal account. Neither should be dismissed merely because it questions the preferred conclusion.

A developed response can judge media influential as agenda setters and channels of accountability while being cautious about their independent electoral effect. It should weigh reach, credibility, competing information and the type of audience. Media can influence parties' strategy and governing behaviour even when no measurable voter changes sides. This broader conception makes between-election influence visible instead of treating the election result as the only outcome that matters.