As the general election campaign drew to a close, opinion polls showed little sign of change, with Labour’s support falling slightly during the campaign.
But after 14 years of Conservative government, Keir Starmer’s Labour Party has consistently polled higher than the Conservatives since the start of 2022.
The Guardian is tracking the latest polling averages from all the major UK pollsters in the run up to election day.
Current voting intentions
Average of polls over a 10-day rolling period, showing UK voting intentions
Changes in voting intentions
Latest average of all polls over 10 days, showing UK voting intentions
The Scottish National Party (SNP) is not included in the data used by The Guardian for the graph above. UK-wide polls place the SNP’s vote share at between 2% and 4% of the national vote share. However, their geographic concentration in Scotland means they will win more seats than other smaller parties with a similar national vote share, such as the Green Party. Scottish-only polls give a much more accurate picture of how the SNP will perform in the next election than the national polls above.
In Britain’s single-member constituency system, opinion polls are limited in their ability to predict who will win: what matters is how many seats each party gets in Parliament, which is decided by individual elections in 650 constituencies.
What the latest polls mean for Congress
Number of seats by political party
Loading…
Chart from The Guardian. Source: Professor Stephen Fisher, Dr John Kenny, Paul Furey, Polina Lijuk, Elections Etc., University of Oxford and University of East Anglia. Projections as of 1 January 1970 and based on averaging forecasts based on voting intentions from opinion polls including uniform change forecasts, multilevel regression and post-stratification (MRP) models and other complex models. Forecasts are averaged within categories before averaging across categories.
Seat forecasts vary, but the ones shown above are averages. We employ three different methods of forecasting how voting intentions will turn into seats: uniform change forecasts, multilevel regression and post-stratification (MRP) models, and other complex models.
Uniform change assumes that national-level variations apply uniformly to all constituencies. The MRP model estimates the relationship between characteristics such as age, gender, and region of residence and which party to vote for, and uses this data to predict the vote in each constituency.
How accurate are the seating predictions?
In Britain’s single-member constituency system, poll numbers don’t clearly correlate with the number of seats they win, because it depends on where the votes are concentrated. Rob Ford, a professor of politics at the University of Manchester, described seat projections from public polls as a “rough rule of thumb,” saying, “Labour might have a 15-point lead that’s not a majority, or a 10-point lead that’s not a majority, depending on where the votes are concentrated.”
When elections are close, polls become less predictive of the outcome. Other limitations of predicting seat numbers from national polls include the fact that it is difficult to infer how many seats the Liberal Democrats will have from national polls; while their national support is much lower than the two major parties, they have a large presence in certain constituencies. Also, national polls are less useful for what will happen in Scotland, and polls there are more rare.
Notes on data
This chart shows the 10-day average trend of support for each party based on polls across the UK, but not Northern Ireland, where the parties are separate. On any given day, the Guardian calculates the average support for each party from all polls published in the past 10 days. Only pollsters that are members of the British Public Opinion Council are included.
YouGov changed its polling methodology from June 4th, but says strategic voting could reduce Labour’s lead.
Seat predictions are provided by Professor Stephen Fisher, Dr John Kenny, Paul Furey, Polina Rizuk, Elections Etc, University of Oxford and University of East Anglia. They are based on averaging predictions based on poll voting intentions including uniform change predictions, multilevel regression and post-stratification (MRP) models and other complex models. Predictions are averaged within categories before averaging across categories. Forecasts are updated twice a week.
Illustrations by Sam Carr. Additional research by Gabrielle Smith, Emma Russell and Lily Smith.
