At the end of May I published a poll on the Newark by-election which put the Conservatives well ahead of UKIP, by 42% to 27%. This raised a few eyebrows, both within parts of the polling fraternity and beyond. Surely the Tory lead was too big? Surely I had missed the UKIP insurgency that would follow their victory in the Euro elections two weeks previously? Surely I was overestimating the number of Tory voters from 2010 who would turn out for the party again despite saying they didn’t know how they would vote?
When the ballots were counted four days after fieldwork ended, each party’s vote share was within the margin of error of my poll. I had been within a point on UKIP, who scored 26% (so much for understating them), two points on Labour, three points on the Liberal Democrats and three points on the Conservatives – whose share was higher, not lower, than in my survey.