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

Baroness Jay of Paddington Excerpts

Tuesday 3rd July 2018 Lords Chamber

[Baroness Jay of Paddington (Lab)]

My Lords, it was a privilege to be a member of such an interesting and thought-provoking committee. I give my thanks to the noble Lord, Lord Lipsey, not just for his excellent chairmanship but for his comprehensive and easy-to-follow description of our report’s findings. I offer my personal warm thanks to the clerk, Sarah Jones, and her team. They gave me enormous support in the few weeks when I was locum chairman and the report was being written, during which the divisions the noble Lord referred to had certainly not gone away.

The noble Lord has well described the sense of crisis about polling which led to this committee being appointed but, at the end of its inquiry, my overall conclusion is that although the polling industry itself may well be facing serious problems, those problems in themselves do not constitute a major threat to our elections or to our democratic process. In other words, I suspect that political polls, particularly voter intention polls, play a much less influential role than many of us have believed. They are likely to be even less significant in future.

In my view there are much greater concerns about some aspects of how digital media affect politics. As your Lordships have heard, although we touched on these in our report we regrettably did not have time to explain them thoroughly. Our emphasis, as the noble Lords, Lord Lipsey and Lord Norton, have shown, was on what we should call mainstream polling. Here, the experts were admirably frank in their admissions of relative failure in the recent headline tests of two general elections and the EU referendum. There is no question but that the UK industry operates to the highest professional standards, but all our witnesses were clear that several factors are today making political polling much more difficult. The noble Lord, Lord Lipsey, has already mentioned a couple and they are worth repeating. There was general agreement that key challenges now include the increasing difficulty of persuading members of the public to take part, the decline in class-based voting and the volatility of the electorate’s choices.

To me, one of the most important factors is the issue of differential turnout, which seems to have been the defining problem in recent experience. As Professor Sir John Curtice, president of the British Polling Council, told us:

“It is pretty clear from the experience of both 2000 and 2017 that estimating correctly who is and who is not going to turn out, particularly the differences in turnout between different demographic groups, is now one of the principal challenges facing the polling industry”.


There seems to be no evidence that any one of these factors or challenges will lessen, let alone disappear, in the foreseeable future. From the polling industry’s perspective, they will probably get worse. Therefore, we should expect that, realistically, national voter intention polls will continue to be uncertain guides to election outcomes, and political polling in general may not be an essential tool in the conduct of our politics. Does this matter? The fundamental question has always been: do polling results directly influence voters’ decisions? Perhaps equally important is the question: do politicians change their actions and policies on the basis of polling results?

I was impressed by how little hard evidence there seems to be that polls persuade individual voters to change their behaviour, but politicians and political parties seem to have been much more susceptible. The noble Lord, Lord Kinnock, summarised this to us by saying:

“The existence of the polls themselves, producing the results that they do”,


all the time, means that,

“the human beings who are leaders cannot be expected to ignore”,

them. I am, after all, the daughter of a Prime Minister, James Callaghan, who apparently postponed a general election on the basis of a private poll. More recently, it is said that polling information affected Gordon Brown’s decision not to hold a snap election in 2007.

History does not yet relate precisely whether the Conservative Party’s apparently unassailable polling position crucially determined the timing of the 2017 election. But in future, perhaps Theresa May and her successors would be well advised to look at the evidence we took and study more carefully the serious difficulties pollsters acknowledge in sampling and predicting today’s electoral behaviour. If I was now a frontline political adviser, I would hesitate a long time before deciding to hold any test of public opinion on the basis of polling—not, I emphasise, because statistical skills are in decline but because social and behavioural changes in the population have created unprecedented challenges in respect of accuracy.

Of course, the importance of the polls, as the noble Lord, Lord Lipsey, said, has often been exaggerated by the way they are reported by the media. Our report rightly includes a great deal of criticism of the headline approach to often complicated statistics and indeed the distortion of nuanced messages. We recommend that the British Polling Council should develop its work on informing journalists about the nature of proper polling and should try to educate editors and producers to translate such things as “margins of error” in less misleading ways. Personally, I am pretty sceptical about how successful such efforts may be. I think it much more likely that the mainstream media will, itself, reduce the prominence it gives to polls, particularly during elections.

My view was reinforced by our evidence from the BBC, ITV and Sky News, all of whom said they had moved away from highlighting polls in their coverage during the 2017 election campaign, while Deborah Mattinson, the founder of Britain Thinks, reported to us that,

“previous experience made a lot of newspapers feel that they had their fingers burned”,

and that,

“there will probably be fewer polls”,

in the future.

Perhaps the best summary of current attitudes was given to us by the Royal Statistical Society. It said that the 2015 election had created a considerable backlash against the polls from both the public and the media, which had been reinforced by the EU referendum and the 2017 election. It concluded that,

“there remains much debate about the usefulness of polls”.

If their usefulness is in doubt, if they have no great impact on individual voting decisions, if they are less prominently reported and if, as a result, our political leaders pay them less attention, surely we should all be less concerned than we have sometimes seemed to be about the impact of polls on our politics and indeed on our democratic process. If we can persuade our political leaders to be less mesmerised, I think we can afford to relax.

As I said earlier, frankly, I am much more troubled about the effect of digital media on politics—the second strand of our committee’s remit. Although it was frustrating, I think that we were right to recognise that this was far too large a topic to be covered properly in our reporting timeframe. However, we took some evidence about the relatively simple issue of the impact of the internet on the mainstream polling industry, where the net clearly makes it cheaper and quicker than before to conduct large surveys. For me, Professor Curtice summarised the present situation crisply by saying that being cheaper and quicker did not necessarily equate to doing polling well.

It was also useful that we succeeded in getting some interesting views on the way that digital news media and social media platforms have revolutionised the way we absorb and use information, and on the relevance of this revolution to our political conduct. We touched on the potential for misinformation in the political sphere, particularly when this can be algorithmically manipulated by those with an interest in malignly influencing any democratic process. The House will of course be aware of the vast number of international investigations into this kind of activity, but, again, we could touch only the surface of the global debate.

None the less, the report has a clear focus on the need for urgent government action to explore possible ways of regulating at least some parts of the digital universe. We emphasised that the use of social media to direct and distort democratic debate adversely was deeply concerning, and we recommended that, particularly with online advertising, the Electoral Commission should be given wider responsibilities. As the noble Lord, Lord Norton, has already said, in their response the Government have not reacted positively to any of the suggestions about increasing the role of the Electoral Commission but they have promoted the extensive role of the digital charter.

When he gave evidence to us, the Secretary of State, Matt Hancock MP, was energetically optimistic and determined to use the charter to find general solutions. The formal government response states:

“The Digital Charter is a rolling programme of work to agree norms and rules for the online world and put them into practice”.


I certainly do not challenge those good intentions but, once again, I am somewhat sceptical about success. When I look at the giant digital corporations based in California and embedded in the American financial and political system, I do not see an easy accommodation with the highly regulated, financially controlled political systems that we are familiar with in western Europe. I am, frankly, not surprised that when a Commons Select Committee travelled to Washington to hold hearings with the west coast corporations—at that time about fake news—the discussion was described by observers as a dialogue of the deaf. It may be that the big platforms themselves will take action, and there are some suggestions that this is beginning to happen; none the less, there needs to be much more international research and many more conversations and negotiations at a technical as well as a political level.

As I said, I regret that our committee was unable to take some of these issues forward. As your Lordships are aware and as the noble Lord, Lord Lipsey, has repeated, we suggested to your Lordships’ Liaison Committee that another committee should be appointed with this specific remit. This proposal was not accepted for the current Session but I very much hope that it will be looked at again.

To conclude, I think that the House of Lords is particularly well placed to conduct a dispassionate and concerted investigation into the broader risks to democracy which threaten us in a world of digitally determined and volatile politics. The Lipsey committee has opened up an enormous agenda which we must return to.