Lords Ashcroft’s Corby poll (and my predictions therefrom…)

by Stephen Tall on October 23, 2012

Today saw the publication by Lord Ashcroft of an opinion poll he’s funded to find out the likely result at the forthcoming parliamentary by-election in Corby: ‘Labour take 22-point lead in Corby’. I tweeted my insta-verdict: it suggests Labour will win Corby by-election battle but lose the general election war. What prompted the judgement was this paragraph from Lord Aschroft’s commentary:

Most Corby voters (56%) are pessimistic about the economy, thinking that in three or four years time things will be no better or even worse than they are now – although the proportion of those who think the right decisions are being made and things will improve in three or four years (44%) has edged up slightly during the campaign. Cameron, Clegg and Osborne are preferred over Miliband and Balls to manage the economy by 48% to 33%, and indeed by three quarters of those who voted Conservative in 2010 but will vote for someone else at the by-election. Though the Tories retain a strong lead on cutting the deficit and the debt, Labour have a small advantage on “getting the economy growing and creating jobs”.

No-one can predict the 2015 general election outcome with any certainty. But my best guess of what will happen is, as I wrote a fortnight ago:

… the fundamentals of the election pitch the Coalition parties will make in 2015 are largely unaffected: Labour landed the UK in an economic mess; we’ve begun clearing it up; we deserve some more time to make it come good. I’m wary of historical analogies, but the next election could easily become another 1992, when mega mid-Parliament Labour poll leads melted away as the economy crawled out of recession and the prospect of a Neil Kinnock-led Labour party winning drove voters back into the arms of the Tories. ‘Hold onto nurse for fear of something worse’ became the unofficial slogan of that Tory campaign, and could easily become so again. It’s safe to vote Labour when the economy’s doing okay (as it was by 1997, and again in 2001 and 2005) but not when it’s teetering (1992, 2010). I’m caricaturing for effect, but I’m sure Labour strategists are aware of the risks that sluggish growth could inflict on their chances in two-and-a-bit years’ time.

So the polling data I’m most interested in at the moment relates to 1) how effective the public thinks the major parties will be at managing the economy in the best interests of them and their families; and 2) who they think are the best team of leaders. On this basis, I think the Tories are in a stronger position right now than Labour. That’s the context to the following Twitter conversation with my old friend Natt on the predictive powers of polling…

The predictive power of polls: a Twitter conversation

Storified by Stephen Tall · Tue, Oct 23 2012 06:47:05

Lord Ashcroft’s poll suggests Labour will win Corby by-election battle but lose the general election war http://bit.ly/PNX8UNStephen Tall
@stephentall Except it doesn’t. It: "would mean a significant overall majority for Ed Miliband if repeated at a general election."Nathaniel Tapley
@Natt Only if you think a by-election swing will be repeated at a general election.Stephen Tall
@stephentall The article considers it likely: "more than 1/4 say they will “almost certainly, or most definitely” not vote C in 2015."Nathaniel Tapley
@stephentall Or, certainly, more likely than a reversion to the results of 2010, which left a hung parliament.Nathaniel Tapley
@Natt I tend to view that forced choice Qs about leadership / econ competence better predictors of GEs than what folk say they’ll do now.Stephen Tall
@stephentall But you think a q about leadership / econ is better predictive of how people people feel about the same q in 3 years’ time?Nathaniel Tapley
@Natt Yep, because it controls for the give ’em a kick response of mid-term polls.Stephen Tall
@stephentall You’re position is "If a general election were held today, but disregarding the results of any elections held today, then…"?Nathaniel Tapley
@Natt Pollsters ask he right Q if you want a snapshot of public opinion. But as predictors of results, are pretty hopeless.Stephen Tall
@stephentall But you deliberately extrapolated a prediction (general election loss) from a poll that made none.Nathaniel Tapley
@natt I gave a view based on my reading of the data. Is that what you mean by "deliberately extrapolated"?Stephen Tall
@stephentall You based your view (or ‘link-title’) on a predictor you think is "pretty hopeless" as a predictor, then?Nathaniel Tapley
@natt No. To re-cap: I think how wd you vote today Q is less useful as predictor than the who you think’s best leader / most econ compet QsStephen Tall
@stephentall But you think the leader / economy q asked 2.5 years before a GE is useful enough to build your own prediction on. I see.Nathaniel Tapley
@stephentall Your analysis, then. "In key marginals, on qs of leadership & econ, Lab doing better than 2010. Suggests worse result in 2015."Nathaniel Tapley
@natt I think they’re the most useful predictive indicators, yes. Best to weight series of factors (incl growth) per Nate Silver of course.Stephen Tall
NB: Nate Silver predicts US election contests using a mix of data, mostly national and local polling but also underlying data, primarily economic conditions: http://en.wikipedia.org/wiki/FiveThirtyEight#2012_U.S._presidential_election
@natt Not worse. Just not as well as they need to do to win. I’d guess/predict a narrow win for Tories in pop vote in 2015 as it stands.Stephen Tall
@stephentall But popular vote has nothing (hyperbole) to do with a general election result. Which is what you were saying the poll revealed.Nathaniel Tapley
@natt My guess is whoever wins pop vote will win most seats. Possible that won’t happen (cf ’51) but not probable.Stephen Tall
@Natt (But, it’s true, I’m more confident saying I think Tories will win pop vote than in saying they’ll win most seats.)Stephen Tall
@stephentall So, when Labour outpolls the Coalition in marginals on economic qs of growth (Q7), you think that shows they will lose a GE?Nathaniel Tapley
You can find the full list of questions and response data here: http://lordashcroftpolls.com/wp-content/uploads/2012/10/Corby-poll-2-full-data-tables.pdf
@natt Q9, you mean? Asks about growth *and* jobs. Actually that’s q messily phrased (Tories tend to win former, Lab latter)…Stephen Tall
@natt …More useful wd be poll Q asking which party best trusted to manage econ for you and your family.Stephen Tall
@stephentall Just chuckling at the wording of part of Q7 "Who do you trust to stop scroungers?"Nathaniel Tapley
@Natt Yeah, 6% says Lib Dems I see *shamed face*Stephen Tall