Our democracy relies on accurate polling. It’s vital we get it right
3 min read
Opinion polling frequently determines the narrative of elections – and may even decide who wins. The industry must do everything it can to ensure accuracy, writes Lord Lipsey
You would expect parliamentarians to be experts on opinion polling. The polls are the best indicator there is of whether their party will be in government. For MPs in marginal seats, indeed, they are the best indicator of whether they will keep their seats or not.
There are some real experts in both houses. The Commons includes Nadhim Zahawi, who co-founded YouGov; in the Lords, Lord Cooper of Windrush was the co-founder of Populus, and Lord Hayward has advised several Conservative grandees on polls. But there are many others whom, you might think, should know more.
I am biased but anyone who wants to learn could turn to the report of a Lords Select Committee which reported this week, entitled ‘The politics of polling’. I chaired the committee after many years of involvement in polling – as polls adviser to James Callaghan, as a commissioner and reporter on polls for newspapers and as Deputy Chair of Full Fact, the fact-checking organisation. It provides answers to many of the questions still floating about concerning polling.
Is polling getting more inaccurate? Well internationally, no, as research by Professor Will Jennings of Southampton University shows. But in Britain, we have cause to worry after three successive failures – the 2015 general election, the Brexit referendum and the 2017 general election.
Polling is undoubtedly getting more difficult. More and more people refuse to be interrogated, so you get a less representative sample. Voting is no longer conditioned mainly by social class, so the old mantra – get the numbers in your sample from each social class right and you won’t end up far out – no longer applies.
Moreover, the small print shows it never was that accurate. Pollsters quote – on quite what basis is unclear – a margin of error of three per cent either way. But that means a poll showing Labour 40 / Conservative 40 might, within the margin of error, represent a six per cent Labour lead or a six per cent Tory lead. That doesn’t tell you very much about who is winning.
This is a select committee report that doesn’t mostly call for government action. The Select Committee wants a bigger role for the industry’s own British Polling Council, for instance by it holding proper open post-mortems after each election. It wants an enhanced role for the official Electoral Commission monitoring voting intention polling conducted and published during elections. But it doesn’t, for example, seek to ban polls in the run-up to elections, as do 16 of the 28 European Union nations. That would be too great an interference with free speech, and might simply trigger off-shore polling to get around the ban.
No reader of our report however, could deny the potential importance of opinion polling in our democracy. It frequently determines the narrative of elections – “coalition of chaos” in 2015, “Tory landslide” in 2017. It thus conditions what the debate is about and may even decide who wins. That is an awesome responsibility indeed – and our report is a call to the polling industry to do everything it can to make sure it gets it as right as it can.
Looking forward, I can see advantage to establishing an all-party committee on psephology to keep parliamentarians abreast of the latest thinking. Any member who would like to join should email email@example.com
Lord Lipsey is chair of the Political Polling and Digital Media Committee.
Get the inside track on what MPs and Peers are talking about. Sign up to The House's morning email for the latest insight and reaction from Parliamentarians, policy-makers and organisations.