Miriam Cates: Have polls replaced principles?

Miriam Cates is a television presenter with GBNews and the former MP for Penistone and Stocksbridge.

Has opinion polling ruined politicians?

Polling – rather than principles – now seems to underpin most policy decisions in Westminster. Is the growing abundance of public opinion polls ruining our politicians’ ability to think for themselves?

The science of polling is nearly 90 years old.

The British Institute of Public Opinion (BIPO) was founded in 1937 by Henry Durant, inspired by the American pollster George Gallup, and the first British poll measured attitudes toward the abdication crisis involving Edward VIII. During the second world war, the Government regularly commissioned polls and surveys to test public opinion on issues such as rationing and conscription. Although we often think of Churchill’s defeat in the 1945 general election as a ‘surprise’, most polls correctly predicted a landslide Labour victory, a success that helped to legitimise the polling industry.

The post-war period saw the rise of commercial polling, and after 2000, the growth of the internet transformed the industry. The days of postal surveys and newspaper phone-ins are long gone. Online panels have replaced many face-to-face interviews, large sample sizes can be collected quickly, and complex statistical modelling is performed in an instant.

Opinion polls are now a constant feature of British politics, increasing politicians’ awareness of their own party’s popularity and the public’s opinions on key issues. It is now possible to see the impacts of policy announcements almost in real time.

In the past, polls were taken with a pinch of salt; pollsters were often wrong in their predictions, including about the outcomes of the 1970 and 1992 general elections. In response to a particularly gloomy prediction about their party’s fate at the next election, MPs could tell journalists in all honesty: “You can’t always trust the polls”. But improved methodology has significantly increased polling accuracy. Although under the British First Past the Post (FPTP) voting system, it will always be challenging to project exactly how many parliamentary seats each party will win, pollsters’ vote share predictions in the last two general elections were broadly correct. And besides, polls are now conducted so frequently that taking an average of the different results – a “poll of polls” – gives a pretty accurate idea of the truth.

Every new day brings a new poll, published and shared on X (formerly Twitter). In the House of Commons tearoom, MPs are now just as likely to pour over YouGov analysis as they are the newspapers, checking their phones for the latest voting intentions like a gambler searching for the horse racing results.

Surely access to more information about what voters think is of benefit to our political class? At worst, isn’t an obsession with checking the latest polls just a harmless habit for MPs waiting to go through the voting lobbies late into the night? Or could an over-supply of opinion polling have wider and negative impacts on our political culture?

Frequent polling causes politicians to live in a constant state of anxiety. When confronted with the latest revelation about their party’s low poll rating, you may hear MPs say: ‘there’s only one poll that counts; that’s the general election’. But this is what the kids call ‘cope’. Being confronted every two or three days with fresh evidence of how likely you are to lose your seat is like being a school pupil who is examined constantly, rather than just at the end of a few years of study. It is very difficult to focus on long-term achievement when you are facing continual assessment.

Constant feedback makes it much more challenging for our elected representatives to hold their nerve. Any Prime Minister who tries to pursue a policy that polls badly will quickly have their wings clipped by MPs. When Rachel Reeves announced the means-testing of the pensioner’s winter fuel allowance soon after the 2024 general election, polling showed the policy to be deeply unpopular, and the Chancellor was soon forced into a U-turn. Non-means-tested spending on pensioners has become completely unaffordable, and asking low-wage young workers to subsidise retired people with reasonable incomes is unjust. Yet because of “public opinion” – as evidenced by polling – pension reform is now in the “too difficult” box for this government.

In presidential systems like France or the United States, leaders seem to be less hamstrung by the polls, especially when, in the case of Trump and Macron, a two-term limit means neither can stand for re-election. But British Prime Ministers are not so fortunate; if they want to govern, they must find support for controversial legislation from backbenchers. When those MPs are now confronted daily by the cold hard reality of the polls, they become less and less likely to hold their nerve and support anything ‘unpopular.’ When U-turns are forced, rebels claim ‘success’ and the government claims it has ‘listened’. And precedents are set.

Another consequence of access to reliable, in-depth polling is that it has become much easier for voters to be segmented. In the late 20th century, pollsters developed caricatures such as ‘Mondeo Man’ or ‘Worcester woman’, but these ‘types’ now seem like blurred images compared to the sharp definition of today’s polling avatars. The polling company More in Common has developed a tool called the ‘seven segments’ of Britain, detailing the characteristics of voters from ‘Progressive Activists’ on the left to ‘Dissenting Disruptors’ on the right. This is a brilliant and important piece of research that is fascinating from a sociological perspective. It allows MPs to see which ages, sexes, income levels, geographic locations, and professions are most likely to vote for them and their policies. Yet the temptation then arises for political parties – like advertising executives – to aim their policies at narrow rather than broader segments of the public.

Last week, Conservative leader Kemi Badenoch promised to reinstate the two-child benefit cap to pay for an increase of £1.6bn to the defence budget. At a time when low birth rates are one of the greatest economic threats, and when by 2028 Britain will spend £36bn a year on state pensions for people who are wealthy enough to be higher rate taxpayers, it seems odd to single out the only benefit aimed at young families for a cut. Odd that is, until you look at the polls. With the Conservatives’ core voting segment – pensioners – bombs poll better than babies.

We see the same principle at play in Labour’s decisions to put VAT on school fees, introduce an ‘anti Muslim hostility Tsar’ or raise the youth minimum wage. Polling shows these policies are popular with particular segments of society. Politics has always been about appealing to voters – how else does one get elected? But when the environment is so data-rich as to allow politicians to pit one section of society against another – rather than making the case for policies in terms of the common good – the result is fragmentation and disintegration.

Can our electoral system cope with this fragmentation? Last month saw the emergence of ‘Restore Britain’, Rupert Lowe’s new party designed to appeal to voters to the right of Reform UK. According to one poll, Restore Britain’s policies on remigration and ethnicity appeal to 10 per cent percent of voters. On the left, a similar fragment supported Jeremy Corbyn’s ‘Your Party’ when it was first announced. Under a system of proportional representation, these splits would come out in the wash, with coalitions forming in parliament after elections as small parties coalesce around broadly left or right-wing agendas. But under the British system, a fragmented political landscape where parties appeal to niche groups guided by polling, could lead to a parliament that is chaotic, impotent and utterly unrepresentative of public opinion. Whoever coined the phrase “let many flowers bloom” had obviously never heard of First Past the Post.

Polling saturation is turning our politicians from leaders into followers. Polls can tell us what voters think about a particular issue, but they cannot tell us if those voters are correct. The public’s views are often contradictory – for example wanting more public spending and lower taxes at the same time – and voters are often poorly informed. One of the driving forces behind the length and strength of covid lockdowns was that polling demonstrated public demand for these policies. The greatest barriers to NHS reforms, welfare cuts and pension reform are the polls. However fashionable it may be to wish for more ‘ordinary people’ to stand for election, we need to be honest about the fact that very few British adults have a sufficiently sophisticated understanding of data, statistics, probability, economics, science, history, philosophy or law to make good judgements on matters of policy and legislation, which is why we pay politicians to devote their time to informing themselves on the public’s behalf. Yet in trying to reflect the polls rather than attempting to persuade the public, too many MPs are trying to outsource their role back to their employer.

In relation to current events, it is noticeable that the main criticism of Farage’s initial stance on the war with Iran is that his position is in contrast with what we know is the majority opinion. Whether Farage is right or wrong is immaterial; the reaction shows that we no longer expect our politicians to lead the public but to follow. Reform UK has since backtracked on its support for the US. Perhaps they have seen the polls.

Polls can also sometimes lead us up the garden path. Although voting intention surveys are simple to interpret, this is not the case with more complex questions of policy. In the ongoing assisted suicide debate, proponents of Kim Leadbeater’s private members’ bill have claimed that the majority of the public support assisted dying. But when more sophisticated polls are conducted, it emerges that most people equate ‘assisted dying’ with palliative care, and support for a state-sponsored suicide falls sharply when the reality of what is proposed it made clear. Polls, like statistics, can be manipulated. Bad data can be worse than no data.

So where do we go from here? How can British political culture benefit from the obvious advantages offered by frequent and accurate polling, without our politicians becoming slaves to public opinion?

I propose a new challenge for both MPs and pollsters. Perhaps polling could be used to measure how effectively the public can be persuaded to change their mind on an unpopular but necessity policies, for example scrapping the triple-lock, or reforming the NHS. The task for politicians – should they choose to accept it – is to set out to educate and inform the public, and convince them of the case for making difficult and painful decisions in the long-term interest of the country. The change in public opinion could be measured at regular intervals, as the problems and solutions become better understood by a greater share of the population. Polling could be used to demonstrate which messages – and which politicians – are the most convincing in persuading voters to change their minds, rather than just reflecting back to them what they already think.

At present, an obsession with frequent polling too often paralyses and disempowers our MPs, and tempts political parties into narrow rather than broad appeals. Yet it might also be possible for polling to be used to reverse the trend for politicians to follow the crowd; perhaps our leaders can learn to lead again.