The YouGov/Sun poll gained an enormous amount of exposure on Sky TV last night when it appeared minutes after the TV debate wrapped up. It put Cameron clearly in the lead and fired up an already excitable Kay Burley. When other polls appeared the tale was somewhat different as ITV/ComRes, the Guardian/ICM and The Mirror all put Clegg first, with The Mirror even reporting that Cameron had come last. So what was going on. A quick look at twitter confirmed a broader sense of astonishment at the YouGov findings. The tweet…
YouGov Poll: Earth round 23% Earth flat 64% Earth other-shaped 13% #LeadersDebate
..was posted by hundreds of voters. A Twitpic that has been viewed over 18,000 times suggests that YouGov has been polling on behalf of either the Tories or Labour to elicit voter fears in the event of the Liberals gaining a big share of the vote.
Twitter posts also pointed to the extraordinary fact that the founder and (until the start of the campaign) CEO of YouGov is a Conservative candidate. It isn’t much of stretch to question the independence and therefore reliability of a polling organisation that may be commercially and politically aligned to one of the major parties. In this game value and reliability are very closely linked.
Ballot Blogs 
A big question is around the rumour that YouGov included “in running” polling in their result – votes cast before all candidates finished speaking. If the poll was genuinely ALL post debate then fair enough, but if any votes were cast before all candidates had finished, it’s neither a “Post debate” poll nor in any way credible.
I’ve asked YouGov via Twitter to confirm that no such votes were counted (http://twitter.com/timprater/statuses/12691414503 – quite recently, so lack of reply so far not unexpected, but Mark Pack also asked them a similar question last night and also seems to have had no reply).
Now had a reply from YouGov, pointing me at http://www.research-live.com/why-the-thirst-to-be-first-can-leave-you-high-and-dry/4002548.blog?sms_ss=twitter
So they DID open the poll early (and I think they suggest 8 minutes early). YouGov claim it was only a handful of votes that were cast in theat early period – but haven’t said how many.
They also claim to have checked the first thousand votes against the second thousand – and not only collected the results, but managed that in the 6 minutes between end of debate and release of poll. They were then satisfied that the two halves matched exactly, despite that being statistically odd in itself (or “no discrepancies were found”, whatever that means).
As the article says:
“In order to be beyond reproach you start fieldwork once the last politician has summed up and sat down”.
Quite.
bullshit…