Skip to main content

Data Update, 7 July 2020


Today we have released an updated version of waves 1-19 of the British Election Study Internet Panel, 2014-2023. This is a routine update intended to fix minor errors and inconsistencies that we and our users have found since the initial release.

The changes are as follows:

  • Renamed “integrityBoris” to “integrityJohnson and “competentBoris” to “competentJohnson” to match style of other leader variables.
  • Deprecated “inequalityGoodBadW3” to the legacy file due to a scripting error that meant that the variable was fielded to the wrong subset of respondents.
  • Fixed “p_religion” to account for changes in YouGov’s data collection process that meant that individuals in the previous version of the data were matched to the wrong religion.
  • Fixed “oslaua” to account for a problem in the previous version of the data where some individuals appeared to live in Northern Ireland or simply had no code, when this was not the case.
  • Relabelled various “p_turnout” variables that had been coded in the previous version as “DNV”/“Voted” to “No, I did not vote”/“Yes, I voted” to match the documentation.
  • Relabelled “p_eurefvote” from “Stay/Remain in the EU” and “Leave the EU” to “I voted to remain” and “I voted to leave”.
  • Removed the “profiles_newspaper2W16/W17″ variables, which were available in the previous version of the data though should have been dropped in favour of the “p_paper_read” variables.
  • Fixed an error introduced in wave 19 where “auth4″ was incorrectly coded with value labels from “auth5”.
  • Fixed an issue with “p_edlevelUniW18″ where some observations had gone missing due to a scripting problem.

Persistent known problems include:

  • For “savingsAmtbW14”, the “£75,000-£99,999” category was incorrectly given the label of “£75,000-£999,999” during data collection