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

Baroness O'Neill of Bengarve Excerpts

Tuesday 3rd July 2018 Lords Chamber

[Baroness O’Neill of Bengarve (CB)]

My Lords, the report of the Select Committee on Political Polling and Digital Media, of which I had the honour to be a member, was intended to address one pretty urgent but relatively well-defined topic and then one less well-defined topic, which to me is probably even more important. The urgent topic was to inquire into why the polling organisations provided estimates which in the event turned out not to be as accurate as had been expected in two general elections and the referendum campaign. That was very well defined. The less well-defined topic concerned the role of digital media in political campaigning. So, the remit was actually quite complicated and the Select Committee rather short-lived. For that reason, I am particularly grateful to our chair—the noble Lord, Lord Lipsey—the noble Baroness, Lady Jay, and the clerks for handling a very complicated set of topics that did not entirely gel.

As the report makes clear, the committee concluded that, in the main, problems of recent political polls were probably not due to deficiencies in the conduct of polls by polling companies. That is solid and reassuring, but it is not a reason for complete satisfaction because we also reported that pollsters were encountering greater reluctance to respond, public confidence in what polls report was declining and there were considerable problems with the use of polling results by parts of the media.

The report’s recommendations address some of these issues. They include greater co-ordination between the industry, the professional body—the British Polling Council—and the Market Research Society, as well as between the Electoral Commission and media regulators. They are measured on proportionate suggestions and it is good to see that the Government are taking them fairly seriously. However, the recommendations do not address the wider issues raised by the spread and power of digital media that bear on political polling. I think that this is because we found the evidence patchy and difficult to assess in the brief time available. Indeed, in some cases witnesses suggested in evidence taken in private that matters were worse than they would, or perhaps could, say in public.

As the topic is vast, I will speak only about a few relevant matters. First, digital media include social media but not all digital media are social media. That is fundamental. Social media content is posted by individuals and controversy arises at two points. The first, better-known issue is that content posted by individuals may mislead or harm. Your Lordships’ House has had considerable opportunities to discuss some of the harm that can be done to individuals by certain uses of social media, such as fraud, cyberbullying, trolling, defamation and many more.

Of course, such action also goes on without the support of digital technologies and is usually criminalised. The difference and the difficulty with content posted on social media by individuals is that it may be posted anonymously, so sanctions are very hard to impose. There is a big debate to be had about the effects of social media use that targets individuals and the limits of arguments for permitting anonymously posted content. Anonymity is often supported with claims that it is needed for whistleblowing. That is incorrect; I think that confidentiality is much more relevant than anonymity to whistleblowing, if you want it to work. The second reason for anonymity is to report news under oppressive regimes. Thank God we are not facing that. This is hardly an argument for permitting anonymity, whatever the communication. The rise of anonymous posting is in itself a social phenomenon about which we need to think intensively and urgently.

The second way in which the use of social media can lead to harm is when posted content is organised to reach some but not others, thereby exerting some control over what individuals receive. Targeted advertising and messages may shape the content that individuals receive and can thereby add or limit content that supports—or, alternatively, seeks to undermine—a given cause. We did not obtain any solid evidence of the extent to which the content that individuals receive has been subject to control or influence. That was one of the big gaps in our evidence. Evidently, if we imagine a wide-open conversation of mankind, we can tell ourselves that the more voices are included, the better—for social life and democracy. However, if the spectrum of choices or positions that are heard is being shaped by other considerations and is often selected to support a cause, or limit support for another cause, then fundamental questions arise about the feasibility of democracy in the age of social media, and now to digital media that are not social media.

There is one more effect that social media have. Social media also monetise the data that individuals supply by using those data to organise and target advertising—by which, of course, the companies secure their revenue. Once again, there are legitimate reasons for concern. There is no reason to suppose that the content that is distributed by social media will secure any even or unbiased distribution of information or evidence to electors. In fact, we have good evidence of the contrary happening, although I think not yet evidence of the scale, the effects or the effectiveness with which this is happening. We just know from some empirical studies that there is uneven distribution of content. These, I think, are reasons why the report could not offer a more systematic account of the effects of digital media, especially social media.

However, digital media go further. Digital media include not just social media but other digital enterprises where the content is not posted by individuals; it is made available by organisations; created, we may say, by organisations; and, indeed, invented by organisations sometimes. Some of these organisations, of course, have clear political purposes, including, very frequently, undisclosed and sometimes malign purposes. It is often hard to detect the source or the allegiance of digital media. Here, a blogger may be indistinguishable from a journalist and probably calls himself or herself a journalist. Here, discipline, let alone credentials, may be wholly absent. Here, there is no editor. Yet we talk about digital media as though they consist of professional journalists who are disciplined by editors who seek to provide reliable content for others.

We talk about digital technologies as if they can be regulated. This may be the deepest of our difficulties. It is often said these days that what we need to do with digital media is to make sure that they are not treated as platforms but as publishers. If they were publishers they would, for example, be subject to the law on defamation, to take one simple example. As platforms, they are not. Nor, of course, are the individuals who post stuff anonymously subject to the law of defamation. This is an extraordinary escape from legal and regulatory discipline. Can it be remedied? Until about a year ago I thought so.

I think we face two major obstacles in addressing what digital media can do. One is the jurisdictional problem. It is extraordinarily easy for these technologies to shift their supposed location: they have very little fixed infrastructure and they can move, as we see by the fact that they pay so little tax. They can move their headquarters where they choose. If we seek to regulate them, it is quite likely that they will find more convenient jurisdictions in which to operate. The other reason why I suspect they cannot be regulated as publishers is that being a publisher is, as many of us know, pretty arduous. You have to read the stuff. There is too much, however, that is posted; they could not carry out the due diligence that is the daily work of publishers.

We talk as if we still lived in a world in which journalism can be reliably distinguished from self-expression, in which political advertising can be identified by seeing who paid the bill. I think that is given the lie by the fact that what we are actually regulating is the paid-for advertising of the political parties during election campaigns, a very narrow form of control when all sorts of other things are going on. I do not think it will prove viable for much longer to regulate only advertising by political parties during election campaigns and to turn a blind eye to all the other advertisers using the same technologies and spreading what they choose to spread. Political persuasion is now cheap and it can be done by those who have no business doing it. We are all aware that the mighty Facebook apparently did not realise that it was hosting political advertisements that had been funded from Russia. I think that is a warning call for all of us. If we are to retain democracy we have to find ways of detecting and ending practices of this sort.