Skip to main content

Data Release: Wave 20 of the British Election Study Internet Panel, 2014-2023

The British Election Study Team

The British Election Study is pleased to announce the release of wave 20 of our Internet Panel survey. Wave 20 constitutes the first wave of the panel to occur after the outbreak of the coronavirus crisis.

31,468 respondents took wave 20 of the British Election Study. 17,794 of these also took wave 19, an overall wave on wave retention rate of 55.3%. 898 respondents took all of the first 20 waves of the survey, 2.9% of respondents who originally took wave 1. Likewise, 2,100 respondents (12.7%) took all waves between wave 9 (the post-EU referendum wave) and wave 20 and 3,885 respondents (19.5%) took all waves between waves 13 (the post-2017 election wave) and wave 20. Wave 20 was conducted by YouGov between 3 and 21 June 2020.

The 2019 British Election Study is managed via a consortium of the University of Manchester and The University of Oxford. The Scientific Leadership Team is comprised of Professors Ed Fieldhouse, Jane Green, Geoff Evans, Jon Mellon, and Chris Prosser. The team is supported by researchers Dr Roosmarijn de Geus and Jack Bailey.

As well as introduce new data, we have also amended some errors to the existing data. These changes are as follows:

  • Fixed p_past_vote_2019 to rely on generalElectionVoteW19 and not only the original profiles data from YouGov as the 2015 and 2017 variables do.
  • Amended small typo in the variable description of the p_paper_read variables that referred to a “newsaper” and not a “newspaper”.
  • Fixed issue which falsely suggested that we had run the partyPriorities_costlive question in W17. This was in fact a question about Brexit and is now coded as partyPriorities_brexitW17 as appropriate.
  • Removed xsexualityW10 in favour of the various “p_sexuality” variables which capture the same information.
  • Fixed an error where selfPriorities_99W17 had the labels “Got a lot worse” and “Got a little worse” instead of “Yes” and “No”.
  • Dropped 477 cases from the data which had no wave indicators and had not responded to any questions. All 477 occurred in either waves 1, 2, or 3. They appear in the pre-processed data we receive from YouGov and were likely included by mistake.
  • Renamed “euRefVotePost” to “euRefVotePostal” to avoid confusion with the new variable “euRefVoteAfter” which measures attitudes towards rejoining or staying out of the EU now that the UK has left the bloc.
  • Renamed knowf2f4W17 to knowf2f7W17 to account for changes to question wording in wave 17, where respondents were asked “The UK uses proportional representation for general elections” and not “The Liberal Democrats favour a system of proportional representation for Westminster elections”.
  • Combined the various national identity items from waves 11 and 12 into a set of variables that span both waves. For example, britishnessW11 and britishnessW12 now become britishnessW11W12. This accounts for the fact that the wave 12 waves were included only as a top-up item.
  • Fixed minor issue in motherCountryBorn and fatherCountryBorn where some countries were coded incorrectly and where those whose parents were born in Ireland were coded as “EU pre-2004” and not “Ireland”.
  • Omitted variable “coalitiontreatW14” which was included by mistake and supposed to be held back for a user request.
  • Changed variable label for “polPreferToFight” from “Politicians more concerned with fighting each other than furthering UK” to “Parties and politicians are more concerned with fighting each other than furthering the public interest”.
  • Fixed subjClassW10W11 and W12W14 to have the same variable labels as the rest of the subjClass variables.
  • Merged voted2015newwordingW15 into p_turnout_2015.
  • Merged discussPolDaysW3 into a new topup variable discussPolDaysW2W3.
  • Added missing variable labels to ns_secW20 and ns_sec_analyticW20 to match existing ns_sec variable labels.
  • Recoded p_household_children so that “5 or more” is now the largest category.

The data also contains the following known problems:

  • savingsAmtbW14: 75,000-99,999 category incorrectly given label of 75,000-999,999.