Explainer

What is Voting Intention? The UK’s Key Political Poll Explained

Voting intention (VI) is the single most important number in British political polling. It answers one question: if a general election were held tomorrow, which party would you vote for?

Every week, polling companies such as YouGov, Ipsos, and Survation ask thousands of UK adults exactly this question. The aggregated results — weighted for age, gender, region, and past vote — produce the headline numbers you see in newspapers and on sites like BritPolls.

What does voting intention actually measure?

VI is not a prediction of an election result. It measures the current state of public opinion — specifically, how voters would behave if forced to choose today. It does not account for turnout, tactical voting, or the effect of a campaign. A party polling at 28% in May 2026 is not certain to win 28% of the vote at a general election in 2029.

In May 2026, UK VI polling shows a dramatically changed landscape from the 2024 General Election result:

• Reform UK: 28% (up from 14.3% at GE2024)
• Labour: 18% (down from 33.7% at GE2024)
• Conservatives: 19% (down from 23.7% at GE2024)
• Lib Dems: 11% (broadly similar to GE2024)
• Greens: 10% (up from 6.7% at GE2024)
• SNP: 3% GB-wide

Understanding why these numbers matter requires understanding how they are produced. See our full guide to how opinion polls work.

How are voting intention polls conducted?

The majority of UK VI polls are conducted online. Respondents are recruited from a pre-recruited online panel and asked a structured set of questions. YouGov — the most prolific UK pollster — recruits respondents from its panel of over two million UK adults.

The standard question wording

The British Polling Council (BPC) requires members to publish their question wording. A typical voting intention question reads: “If there were a UK general election tomorrow, which party would you vote for?” followed by a prompted list of party names. Unprompted and prompted versions produce slightly different results.

Weighting and methodology

Raw survey data is rarely representative. Pollsters apply statistical weights to correct for imbalances. Common weighting variables include age, gender, region, education level, and crucially, 2024 General Election vote recall. Past vote weighting is particularly important and methodologically sensitive — it is the main reason different pollsters using different weighting schemes produce different results (known as house effects).

How often are VI polls published?

• YouGov: Weekly, sometimes twice-weekly. The most frequent publisher
• Ipsos: Monthly. Uses telephone interviewing alongside online
• Survation: Roughly weekly or fortnightly
• Redfield & Wilton: Multiple times per week — the most frequent publisher after YouGov
• Deltapoll: Monthly or fortnightly
• Opinium: Weekly for The Observer

Why do different polls give different results?

Variation between polls — called “house effects” — is normal and expected. Different pollsters use different panels, weighting schemes, and question wording. Always use a polling average across multiple pollsters rather than any single poll. Our voting intention tracker aggregates data from all major UK pollsters.

What is the margin of error?

For a sample of 1,000 respondents, the theoretical margin of error at 95% confidence is approximately ±3 percentage points. A poll showing Labour at 18% and Reform at 28% should be read with that uncertainty in mind — though the 10-point gap between them is well outside the margin of error and represents a real difference.

The British Polling Council sets standards for how polls must be reported and requires members to publish full methodological details.

Voting intention vs. seat projections

Voting intention numbers tell you about vote share; they do not directly tell you about seat distribution. Under First Past the Post, vote shares translate into seat counts in highly non-linear ways. A party at 28% nationally may win far fewer or far more seats than their share implies, depending on how their vote is geographically concentrated.

MRP models are the tool for converting VI data into seat projections. Our GE 2029 forecast uses published MRP data to show what current polls imply for the next election.

How to read a voting intention poll intelligently

• Never treat one poll in isolation — look at the trend across multiple pollsters
• Check the fieldwork dates: a poll completed last week is more current than one from last month
• Note sample sizes: smaller samples mean wider uncertainty
• Understand house effects: know that Kantar tends to show higher Reform numbers than YouGov
• Do not over-interpret small movements: a 2-point change in a single poll is likely noise