Politics

Election 2019: a guide to what the polls mean – and what they don't


Politicians love to hate them, until they show a result that suits. Journalists caution against reading too much into them, but can easily be tempted to take them seriously.

No wonder, then, that the voting public can easily be bemused by the endless supply of election opinion polls: that appear to show one thing and yet have little clear bearing on the final result.

As the general election campaign gets under way, and with the so-called “horse race” attracting enormous interest, the Guardian is launching its “polling station” coverage – not adding to the noise that the polls produce, but helping you pick out the parts that are useful.

What do the polls actually measure?

Pollsters ask the public a range of questions, but in the end one counts above all, traditionally: “If there were a general election tomorrow, which party would you vote for?”

Their samples are representative of the population but cover Great Britain only – with Northern Ireland excluded because none of the principal Westminster parties is competitive there.

Individual polls can vary by five points or more, so it can be more helpful to follow the averages. The Guardian poll tracker, which tracks all national polls over a rolling 14-day period weighting each equally, had, as of 8 November, the Conservatives way out in front on 39%, ahead of Labour on 27%. The Liberal Democrats are on 16% and the Brexit party on 9%.

Compare that with the vote shares in the 2017 general election. The Conservatives achieved 42.4%, four points ahead of where they are now, but Labour’s relative position today is far worse. Last time, the party climbed to 40%, 13 points ahead of its current position.

poll tracker graphic

Both benefited from what appears now to have been a brief and anomalous return to two-party national politics: in 2017, the post-coalition Lib Dems were on just 7.4% and Ukip (the closest to the Brexit party today) 1.8%.

But what does that mean for the actual election result?

Translating poll figures into how many parliamentary seats will be won is notoriously tricky, and polling firms are wary of taking that task on themselves. Psephologists, arguably braver individuals, do, however, try.

One of the better known publicly available models, Electoral Calculus, estimates that on current averages, the Conservatives will win 370 seats, and a sweeping majority of 90, with Jeremy Corbyn’s Labour a distant second on 186.

Electoral Calculus graphic

It is a Tory lead that has increased from 30 October, when it was clear that an election was taking place. At that point, the site was predicting a Tory majority of 76. And the current prediction still allows for the Scottish National party to increase its position in Scotland, taking 48 seats, up from the 35 it achieved in 2017.

What about the dynamics of the campaign?

Pundits argue endlessly about which events in a campaign cut through to voters, although the traditional view had largely been that election campaigns make surprisingly little difference to the final outcome.

That theory was shattered by the 2017 campaign, when Corbyn’s Labour came from 21 points behind to finish with a deficit of just 2.4%. That gap, of course, produced the hung parliament.

It is debatable too, who has had the worst start this time. Jacob Rees-Mogg’s callous initial remarks about the Grenfell Tower victims and the enforced resignation of Alun Cairns as Welsh secretary have been cancelled out by Tom Watson’s departure and the sight of former Labour MP Ian Austin urging voters to back Boris Johnson.

Polling averages also lag events by a couple of weeks, so any genuine movement in sentiment takes a while to show. So in this campaign, the trend lines have been a little different: the Conservatives are up three points, Labour up two. Losing out are the Lib Dems, down two and the Brexit party, down one. Other movements and roundings account for the remaining differences.

So a little over a week into the campaign, there is one tentative trend: a shift in support towards the two big parties, which is yet to be sustained for long enough to count as a consequential shift. Above all, at least at this point, there is an election-winning Tory lead.

What about constituency polls – is there not a different story?

Although theoretically attractive, polling in individual constituencies is not only expensive to conduct but notoriously unreliable. Nonetheless, the Lib Dems commissioned a string of polls from Survation in an attempt to show their position as positively as possible.

They purport to show the Lib Dems leading in Labour held Cambridge and Portsmouth South – with a 15- or 16-point swing from Corbyn’s party – and taking a seat from the Tories in South Cambridgeshire, although that was where Heidi Allen had been expected to stand for the party.

But pollsters such as Martin Boon, who founded Deltapoll, warn that there is a “disconnect between these results and the national picture” because they suggest the Lib Dems doing far better locally that national projections suggest.

What about regional polls – can they be more revealing?

Regional polls can be particularly helpful in capturing variations at an easily understood level and are worth keeping an eye on. Two interesting regional polls from YouGov (a firm that is showing a big Tory lead nationally) have already emerged.

London voting intentions

The first, in London, records that Labour still leads with 39%, but down 16 points on the 2017 general election; the Conservatives are at 29%, down four points, and the Lib Dems up 10 points to 19%, potentially helping Jo Swinson’s party take seats in south-west London. Labour gains in 2017 such as Battersea and Kensington are at risk, and the party misses out on key seats it will have hoped to gain.

Voting intentions in Wales

Finally, a poll in Wales for ITV showed that Labour had staged a mini-recovery to retake the lead, at 29%, making a gain of four points from mid-October. But Corbyn’s party is still 20 points down from its 2017 general election result there, and Prof Roger Awan-Scully has forecast that Labour would lose 10 seats out of 28 held, with the Conservatives picking up nine and the Lib Dems one.

So far, the regional picture mirrors the national one.



READ SOURCE

Leave a Reply

This website uses cookies. By continuing to use this site, you accept our use of cookies.