On Saturday 4th April ComRes confirmed conducting the South Thanet poll which was reported to have been leaked to the Mail on Sunday. The poll was commissioned by Chartwell Political. We have now also published the full tables.
There have been a number of incorrect reports of the methodology used for this poll which this statement aims to clarify.
UKIP were reported to have said that the poll's headline figures understated support for Nigel Farage 'because it included people who were not likely to vote and who tended to lean towards the traditional parties'. The tables prove that both assertions are plainly wrong.
As is standard methodology for all ComRes voting intention polls, respondents were asked their likelihood to vote on a scale of 1-10. Only those likely to vote (answering 5/10 or higher) were then asked their voting intention and those answers were weighted according to their likelihood to vote (ie. 0.5 for those saying 5/10, 0.6 for 6/10 up to 1 for those 10/10 certain to vote). The tables clearly indicate that people who were not likely to vote were not included in the final voting intention figures.
Those likely to vote but say they do not know who they would vote for were asked a supplementary "squeeze" question to understand which way they were likely to lean. Their answers were then weighted by their likelihood to vote.
As is standard for ComRes polls of individual constituencies, no past vote weight was applied due to the inevitable churn of residents in and out of a single constituency since the last general election.
The names of all the candidates known at the time were used in the wording of the voting intention question.
The poll was conducted to the highest standards and ComRes stands firmly by its methodology.