The Politics of Polling (Political Polling and Digital Media Committee Report)

Lord Smith of Hindhead Excerpts

Tuesday 3rd July 2018 Lords Chamber

[Lord Smith of Hindhead (Con)]

My Lords, it was an honour to serve on this Select Committee under the chairmanship of the noble Lord, Lord Lipsey. It is good to see him back to full strength following his illness towards the end of the committee’s deliberations.

Recent political polling, as has already been mentioned, has been problematic, with three of the last seven elections being incorrectly called. The general election in 2015 was widely considered to be an embarrassment for the polling industry, and this poor standing was not enhanced by its performance at the 2017 general election and the referendum. What became clear to me during the process of taking evidence on polling was a certain amount of blame-gaming between the polling companies and the media. Polls taken during a campaign are effectively a snapshot of a horserace. There is no real way in which these snapshots can be tested, since the election has not yet taken place; and while the polling companies do state that there are margins of error and that often the sample sizes are not significant, these health warnings do not, or rarely, translate into how a poll is reported in the media.

As a result, when the polls have been proved to be incorrect following the finishing post of an election, the media will blame the pollsters for getting it wrong, and the pollsters will blame the media for not properly reporting the polls that have been published. In fairness to both the polling companies and the media, there is a recognition of these shortcomings. I hope that during future elections there will be improved accuracy of polls, with polling companies recognising that people’s social habits have changed, and that as a result, the methods used for taking polls will be updated or, perhaps more interestingly, we will revert to the old style of knocking on doors. The media need to take more care in reporting polls, making sure that sample sizes and margins of error are clearly indicated and explained.

What is the point of polling? That is a question that I frequently asked—a question that on several occasions proved to be the most difficult one for witnesses to answer. In fact, I am still not sure that some of them have worked out what the answer should be. Many witnesses referred to the fact that the exit polls had been very accurate in recent elections, but that should not really be such a surprise: exit polls are not a snapshot of the horserace but, of course, a shot of the finishing line, when people have voted, and they are taken from a much larger and UK-wide sample. I also asked, rather waspishly, what was the point of the exit poll, apart from giving Mr Dimbleby and others something to talk about from 10 pm on election night until 6 am when the outcome is known. Again, there seemed to be some hesitation, before a falling back on the need to educate and entertain.

However, the question to which I found it most difficult to establish a definitive answer was whether polls have an effect on the way that people vote. In other words, could the outcome of future elections be affected by misleading information, potentially distorting the democratic process? My own view is that some people like to be on the winning side, and could therefore be swayed by a poll indicating a particular party’s strength, while others might be persuaded to vote against a party, if a poll were to indicate a landslide, in order to create a balance. I should point out that this is merely my gut instinct: it was shared by several witnesses but not all. What was clear, however, was that there is no grand conspiracy involving polls being deliberately wrong or deliberately misreported to create such a hypothetical scenario.

I cannot express how important I think it is that our recommendation for a proper and enforceable framework for the regulation of polling be put in place. This is the challenge for the British Polling Council, media regulators and the Electoral Commission, since if polling continues to be inaccurate then there is a risk that future elections could be affected by misleading information, potentially distorting the democratic process.

The question of the effect of digital media on polling was perhaps the most interesting part of the Committee’s work, but as our report states, the issues raised were so significant that they stretched beyond our remit and we did not have the capacity to give due attention to them all, although I am very pleased that the noble Baroness, Lady O’Neill, covered many of the points in her excellent address. I do, however, know a lot more about manned and unmanned bots than I did previously and there is concern, rightly, that measures should be put in place to tackle online manipulation and disinformation.

I am pleased that our recommendation to co-ordinate a strong international response to tackle any attempt to maliciously interfere with the UK’s democratic process has been supported by the Government and that work is already under way with allies and partners to promote a collective response. This is the real challenge for the future.