Why it’s time to stop obsessing about opinion polls' wild predictions

As a native of Argyll, I would be first to propound its many virtues. Alas, one of these has never been to elect a Labour MP. If a major opinion poll is to be believed, this is about to change.

I will keep as a collector’s item the Daily Telegraph map which turns Argyll red, from the Mull of Kintyre to the sands of Tiree. It is an impressive sight and victory will be a massive advance on the 6.8 per cent vote share achieved by Labour’s candidate in 2019.

True, the seat now takes in South Lochaber and Ardnamurchan. Over the decades, the Argyll constituency stretched into Bute and Helensburgh, so who knows? I hope the Telegraph map is indeed the harbinger of historic breakthrough but will not crack open my best bottle of Bunnahabhain just yet. That opinion poll by Savanta was reported in the Telegraph under the massive headline “SNP wipeout” and predicted that our previously predominant patriots will be reduced to just eight MPs. Sadly, I don’t believe a word of it.

It's real votes that matter, not pollsters' attempts to predict them (Picture: Matt Cardy/Getty Images)

Southern editions of the Telegraph had the headline “Tory wipeout” and Savanta foresaw Sunak’s shower returning 53 MPs, the great leader not among them. YouGov for the Times predicted 108 Tories and 20 from the SNP. Place your bets, as they say in Downing Street. The Telegraph reported proudly: “The analysis comes from a method called Multilevel Regression and Post-stratification, or MRP, which allows pollsters to take survey results and predict results in individual seats.” Well, up to a point, Lord Copper.

Polls are useful indicators of the broad direction of political travel while real votes are the ones that matter. I am more impressed for example by Labour’s success in a council by-election in Clydebank last week with a 12 per cent swing from the SNP, than in the wonders of Multilevel Regression and Post-stratification.

The danger of polls, particularly in the run-up to an election, is that they influence people not only in how they should vote but whether it is necessary to vote at all. If “wipeouts” are assured, is it really worth the effort of making one’s way to the polling station on a wet Thursday evening?

Exaggerated predictions of “wipeout” also invite the sympathy card which is just about the only one the Tories have left to play. Please vote for us, they beg, to prevent a huge Labour majority. Democracy needs us an opposition!

After 14 years, it is quite pathetic that this is the best they can offer. But what if the polls are indeed exaggerated and sympathy for Sunak and the Tories aids survival rather than damage limitation? Best to beware unintended consequences.

In Scotland, there is a particular need to ignore polls because there are so many complicating factors. For starters, the SNP’s core vote is still stronger than what the Tories are reduced to in England. Heaven knows why, given their record, but it is the reality.

Labour government will not lift UK block on Scottish gender reforms

Only two Scottish polls since 2019 have put them below 30 per cent and that is unlikely to change before July 4. By contrast, the Tories are down in the low 20s and falling, with Reform nipping at their heels. It is difficult to reconcile one prospective “wipeout” with the other.

All sorts of factors will influence outcomes. In some Scottish seats Labour is predicted to take, the majorities to overcome are huge. Some of this was due to the Corbyn factor which depressed the Labour vote. When there was no prospect of forming a government, the choice was between two options for opposition. That is now very different.

In about 40 seats, the SNP have the starting point of incumbency. Most of its MPs have proved themselves comprehensively useless but a few are better than that and may get the benefit of the doubt. Uniform swings will not apply.

Then there is the redrawing of constituency boundaries which means past form becomes an uncertain guide. And all this is before we come to tactical voting which will influence large numbers of votes – particularly where there is a consensus that Scotland has had quite enough of the SNP as either a government at Holyrood or opposition at Westminster.

SNP facing investigation over claim of 'misuse' of election mail stamps

Put all this together and, even without the benefit of training in Multilevel Regression and Post Stratification, my nose tells me that many – perhaps most – Scottish seats will be decided by a few percentage points which means every vote really will count, whatever opinion polls say.

There is a case for banning them in the last couple of weeks before an election and I heard my old friend George Foulkes arguing passionately in support of looking at how this works in other parliamentary democracies. In the meantime, an even stronger antidote is to ignore them – and vote according to either belief or pragmatism.

The likely outcome – and I will not fall into my own trap by accepting the polls as guarantor – is a Labour government. A big question for Scottish voters is whether we will be better served through a strong presence within that government of Scottish Labour MPs, as opposed to the brand of opposition the nationalists offer – denouncing everyone and influencing nothing.

Devolved issues as well as reserved ones have rightly featured in the campaign and it has become pretty clear that patience has worn thin with the SNP’s performance in government and the stock excuses offered, with everyone to blame but themselves for multiple failings.

It is entirely proper that this record should be judged on July 4, for it can kick-start the case for Scotland to have two like-minded governments, working together for transformational outcomes rather than driven by a permanent search for division and evasion of responsibility.

Meanwhile, I will await the Argyll result with particular interest!