When I reported yesterday on Tory claims that a rerun of Lord Ashcroft's marginals poll (with the sitting MPs named) put them two points ahead of Labour (rather than 14 points behind), I wrote that the good Lord would be "watching closely". As anticipated, Ashcroft has now intervened, accusing the Tories, in a withering piece on ConservativeHome, of resorting to "comfort polling".
In the original article, by the Telegraph's Dan Hodges, a "Tory analyst" was quoted as saying: "We reran it in the seats we hold but included the name of the sitting MP. We were ahead by 2 per cent." The existence of an "incumbency effect" (which naming the existing MP would encourage) is not disputed. In 2010, both Tory and Labour incumbents performed disproportionately well. Labour's vote fell by 5.2 per cent in those seats where the incumbent stood again, compared to 7.4 per cent elsewhere, while the Tories' rose by 4.1 per cent in incumbent seats, compared to 2.9 per cent elsewhere.
But as I pointed out, this leaves some important questions unanswered. Were other candidates named? (Voters may well prefer an individual to an amorphous party.) How were the questions worded? ("Will you vote for your hard-working local MP Matthew Hancock rather than the 'welfare party'?") What was the sample size? What was the weighting?
These concerns are also raised by Ashcroft, who writes: "As worrying is the idea that the private poll 'included the name of the sitting MP' but – by implication, and of necessity since not all candidates will have been selected – nobody else. It would not be surprising if this skewed the result considerably in favour of the incumbent. Did the party really spend money on a such a flawed survey?
"Even if they did, it sounds unlikely that you could transform a double-digit deficit to a two-point lead simply by naming the MP. Which makes one wonder about other aspects of the poll. Which seats was it conducted in? What was the sample size? How was it weighted? How were the questions worded? With most polls, including mine, all this information is published. Where it is not, it is worth taking any reported results with more than a pinch of salt." He also asked, "why haven't the results been published per the pollsters code".
As Ashcroft notes, off-the-record briefings on private polling should always be treated with scepticism. With this in mind, I've written to the British Polling Council asking whether the poll should now be published in accordance with its disclosure rules. As the BPC states, "In the event that the results of a privately commissioned poll are made public by a third party (i.e. external to the organisation that commissioned the survey, its employees and its agents — for example the leak of embargoed research) the survey organisation must place information on its website within two working days in order to place the information that has been released into proper context."
Andrew Cooper, David Cameron's former director of strategy and the founder of Populus (and co-founder of the BPC), has argued on Twitter that the poll should be released (if it exists at all).
— Andrew Cooper (@ATCooperNo10) November 21, 2013
The outcome is likely to depend on whether the pollster in question is a member of the BPC. Since surveys must be released "within two working days", I should get an answer shortly.