Polling methodology
Labour Together conducts its polling online, using the same process as most other pollsters in the UK.
The data is collected from two online panels using quota sampling (further details below) and then weighted to be representative of the GB adult population. Responses are weighted by age, gender, region, education level, vote in the 2019 general election, vote in the 2016 EU referendum and the amount of attention paid to politics. Targets for the UK population are gathered from the ONS, along with the results of the 2019 general election and 2016 EU referendum, the British Election Study face-to-face survey, and the annual population survey.
Headline voting intention is calculated by excluding those who say they are undecided or will not vote, and then adjusting weighting according to the respondents’ likelihood to vote.
Methodology and panel providers may differ in polls of specific audiences, and this will be made clear in published data tables.
Sample composition and weighting
Labour Together conducts polling by recruiting respondents from various panels (groups of people who have signed up to do surveys).
Respondents are selected for final inclusion in the sample according to their demographic profile, in line with quotas on age, gender, region, education level, vote in the 2019 General Election and the amount of attention paid to politics.
Sample quality is ensured by the use of captcha, and the removal of those who have a total response time that is less than two standard deviations below that of the mean, or who fail in-built ‘straighliner’ and consistency checks.
The final sample is weighted using random iterative method weighting to account for any over or under representation of particular groups within the final sample.