Those UKIP Polls of Swing Seats

John Rentoul

bown Those UKIP Polls of Swing SeatsI still haven’t had a chance to look closely at the Survation polls of marginal seats carried out for Alan Bown, a UKIP donor. I commented on them last week, and today Survation has published polls in four more constituencies, while Bown has taken out a full-page advertisement in The Daily Telegraph (pictured) making the extravagant claim that the polls show UKIP “poised for a breakthrough” at the general election.

However, a Conservative source points out that the polls seem to flatter UKIP in four ways.

1. The samples are not politically-adjusted.

As Anthony Wells wrote, the poll in Thanet South seemed to have too high a proportion of people who said they had voted Labour at the last election:

There are some methodological changes from Survation’s past constituency polls. Previously they’ve weighted constituency polls by 2010 past vote and reallocated don’t knows based on past vote, in the same way they do for their national polls (though for practical reasons they do national polls online, but local polls by phone). For the latest polls they’ve changed method – no longer using political weighting, and not reallocating don’t knows.

2. The don’t knows and won’t says are high.

There are unusually high levels of people who say that they are “certain to vote” (10 out of 10 on a “likelihood to vote” scale) but who don’t know or won’t say how they will vote. In South Thanet 21% of those who are certain to vote fail to give a voting intention. In Great Grimsby 31% of those who say they would vote don’t say for which party; in Dudley North, 38%.

3. The don’t knows are not re-allocated.

My source says that Survation’s usual practice of re-allocating don’t knows to the party for which they say they voted in 2010 has not been followed in these polls. That would obviously suppress the main parties’ votes to the advantage of UKIP.

4. Survation prompts for UKIP.

Survation does this for its national polls, which is one reason why it gives a high UKIP share in those polls too. There is a debate among pollsters as to whether UKIP should be mentioned in the first question about how respondents intend to vote. Most pollsters mention Labour, Conservative and Lib Dem (and the SNP or Plaid Cymru in Scotland and Wales) or “some other party”. Only if the respondent says “other” does the questionnaire go to the full list.

Peter Kellner, president of YouGov, points to a lot of work his company has done that suggests that, although this might seem unfair to UKIP, the standard question format produces a more accurate result.

I think Kellner is more likely to be right than Survation on this point, but, even if he is not, the Conservatives have a good argument that there are four unusual features of these polls, all of which tend to inflate the UKIP share of the vote at the expense of the Conservatives in particular.

Update: Patrick Briône, Director of Research at Survation, has responded persuasively to these criticisms at Political Betting.

Tagged in: , ,

Property search
Browse by area

Latest from Independent journalists on Twitter