Should we ban opinion polls from being published in the lead-up to election day?

by Stephen Tall on November 13, 2013

One-third of MPs (including a third of Lib Dem MPs) say yes – but more are opposed. At least that’s the finding of a ComRes survey of 159 MPs in the wake of the Indian Election Commission banning exit polls in the five states holding elections this month, plus a ban on any opinion polls in the final 48 hours of campaigning.

    Would you support or oppose a ban on the publication of opinion polls for a defined period prior to General Elections?

    Support
    All 30%

    Con 25%
    Lab 35%
    LibD 32%

    Oppose
    All 45%

    Con 49%
    Lab 39%
    LibD 38%

    Don’t know
    All 25%

    Con 26%
    Lab 26%
    LibD 30%

Here’s what Andrew Hawkins, chairman of ComRes, has to say about it:

“The internet has had two profound effects on political polling. Firstly it has made it quicker and cheaper, but it is too soon to tell yet whether it is as accurate as the telephone polling. Secondly the internet, and especially the advance of social and other online media, renders entirely nugatory any attempt to ban the publication of opinion polls in the run-up to elections.

“Of the polls conducted during the 2010 election campaign, just four were face-to-face, 33 were by telephone and 51 online – and the final telephone polls had a better accuracy record than the online ones. Interestingly though, as Nate Silver points out, the average error in the 2012 Presidential Race was smaller among online pollsters than it was for telephone. The 2015 UK General Election will doubtless help decide which methodology is likely to dominate in the longer term.”

Many countries operate an ‘election silence’ – a ban on campaigning which often includes a ban on publishing opinion polls.

Though I have an instinctive dislike of banning things, part of me’s attracted to the idea. It would introduce a genuine uncertainty into the result which might, just might, boost turnout. After all predictions of an election landslide can become self-fulfilling – the lowest general election turnout under the mass franchise was in 2001, at least in part because the result was never seriously in doubt.

However, pragmatically, I suspect what Andrew Hawkins says is right: that there is little hope of preventing leaks of polls online, some of which may even be semi-authorised by parties or interest groups looking to manage expectations or even to influence the result. Better to have the full information published for all to see.

Incidentally, Lib Dem MPs are most likely to trust polls conducted face-to-face (95%) rather than by telephone (85%) – just 49% trust internet polls. In reality almost all political polls are conducted by phone or internet and how the poll is conducted is one of many factors – and probably not the most important. As Anthony Wells pointed out here last year, how pollsters account for likelihood to vote, deal with don’t knows, weight their samples and prompt voting intention all affect the numbers that get reported.

(NB: ComRes interviewed a representative sample of 159 MPs online and using self-completion questionnaire in April-May 2013. Data were weighted to reflect the House of Commons in terms of party and constituency region. So I’d guess 15-16 Lib Dem MPs replied.)

* Stephen Tall is Co-Editor of Liberal Democrat Voice, and editor of the 2013 publication, The Coalition and Beyond: Liberal Reforms for the Decade Ahead. He is also a Research Associate for the liberal think-tank CentreForum and writes at his own site, The Collected Stephen Tall.