How political opinion polls are warping democracy as Scottish Parliament election looms
Modern political polls use a complicated methodology that does not always reflect the reality on the ground and they can have an effect on the way people vote
Reading the media, you could be forgiven for thinking that May’s Scottish elections are a vote on the UK Government or Keir Starmer’s leadership. That reading is wide of the mark.
Starmer is, in all likelihood, going nowhere, whatever the result, and UK Labour has well over half its term left in office. What this framing does achieve is something else: it blurs the line between Holyrood and Westminster, and diverts attention from Scotland’s growing domestic challenges and the SNP’s failures and scandals.
This sleight of hand is undoubtedly exemplified by how we now treat polling. I served as an MP from the late 70s to the early 2000s, and the change since then has been striking. Polling once meant face-to-face interviews and landline calls that reached a genuinely mixed cross-section of the public.
The adherence to either random or quota sampling and the human connection made it a more faithful reflection of the national mood.
Hamilton by-election a case in point
Today, much of that has been replaced by online panels made up of regular respondents who are far from random. To try to compensate, pollsters lean heavily on complex weighting systems designed to make their samples resemble the electorate.
But this makes the result something of a black box, where methodology, question design and assumptions can effectively shape the outcome as much as the answers themselves.
Experience suggests this model is not capturing political reality. In the Hamilton by-election last year, the consensus among pollsters and commentators was that Labour was out of contention, with most predicting a straight fight between a struggling SNP and a rising Reform, with some placing Labour a distant third.
However, when the real votes were counted, Scottish Labour won decisively. It was not just a missed margin; it was a failure to grasp the direction of travel. Voters in places like Hamilton are not abstract data points. They are making judgments about who can deliver in their communities.
‘Shy Tories’
Respected analysts such as Professor John Curtice acknowledge that a poll should be treated as snapshot, not a forecast, and there are clear technical limits. Most polls carry a margin of error of at least plus or minus three per cent, which in a fragmented contest is more than enough to distort the result. In a charged political climate, some voters are reluctant to state their true preference.
In Scotland, these voters were often ‘shy Tories’ and there’s a widespread assumption they will now be ‘shy Reform’. But I’m not convinced a huge swing to Reform, which much of the press are predicting, will come to fruition.
I don’t believe that they could leapfrog Labour and the Tories in Scotland, particularly led by a multi-millionaire and with no real record of service on the ground.
There is also a distortion in how parties themselves deploy polling. National vote share figures are routinely repackaged into local scare stories, with parties presenting themselves as the only viable bulwark against Reform.
This is dishonest, but it can become a self-fulfilling logic: voters are told that only one party can win, and some duly fall in behind it, out of calculation, not conviction.
Spectre of a Reform surge
We saw shades of this in the Gorton by-election, where the Greens leaned heavily on the spectre of a Reform surge. In reality, the combined progressive vote remained significantly larger, and Labour was not on the brink of defeat until that narrative took hold. Polling, in this sense, does not just measure opinion. It can begin to shape it.
This isn’t just a theory. Following the 2015 general election polling failure, the London School of Economics’ Professor Jouni Kuha concluded that polls were not independent of the political process and can actively shift voter expectations and, as a result, their behaviour.
If voting for a particular party is framed as a “lost cause”, it encourages some voters, who would have backed that party, to engage in unnecessary tactical voting.
A more representative electoral system, such as Scotland’s, can mitigate this, but it remains a very real risk. Many mature democracies, including Canada, France, and Norway, operate a ‘blackout period’, when polling publication is restricted in the final days of a campaign, to specifically get around this type of interference.
In the UK we have a brief ‘window of silence’ – on election day itself – but this hardly allows for a proper cooling-off period, which could force the media to pivot from the ‘horse-race’ numbers back to policy debate.
Time for a ban?
Whether or not some form of ban is desirable, it’s undeniable that we’re hooked on polling – we are hardwired to be drawn to patterns and predictions – so it’s high time we implemented some form of effective regulation.
After a Lords inquiry eight or so years ago, which warned the government may have to create a regulatory body for polling, similar to Ofcom, the British Polling Council fended off this threat by committing to stronger self-regulation. But I’ve yet to see any evidence of this commitment coming into effect, beyond slightly stricter adherence to model transparency.
That’s great, but the average voter has neither the time nor the inclination to study the specific variables behind a poll. The media can’t always be relied upon, partly because many outlets seek out polls which serve a narrative, but also because understanding multi-level regression with post-stratification is incredibly difficult, unless you’re a data scientist.
We need an enforceable system to oversee the conduct and publication of polls to ensure they’re not used for partisan manipulation, and until the BPC takes firm action the public should treat every ‘major swing’ with the scepticism it deserves. Ultimately, the only reliable poll is an election and I will be studiously ignoring every prediction, up until May 7.
George Foulkes is a life peer and a former Labour MP and minister
Comments
Want to join the conversation? Please or to comment on this article.