“Other people have good ideas”

I attended a lecture last week, it was advertised by its conveners on twitter. The lecture was videoed and I am expecting the video to be posted on Youtube. I’ve made some notes, some about what the lecturers said and some about the thoughts they provoked.  I try to offer some value on this blog, however much of this article is reporting the views of the three lecturers; I hope there’s enough of what I say to be worth the read.

The key speaker was Professor Kate Dommett (X:@katedommett) of Sheffield University. She has written a book with a similar title, and I have found this page, although I am not sure if it’s a review or an advertisement. However, the book seems well sourced, and covers numerous geographies.

We ask if Data driven campaigning is to be feared? Recent interest about the behaviour of political data scientists is almost certainly sparked by Carroll Cadwalladyr’s work exposing both Facebook and Cambridge Analytica’s criminal behaviour with respect to privacy during the referendum campaign. Professor Dommett references the ICO report into the affair. I have covered the affair in my blog [and on Medium] and also made mirrors of critical articles published by Cadwalladyr in the Guardian.

From the ICO’s report to Parliament,

We sent 11 warning letters requiring action by the main political parties, backed by our intention to issue assessment notices for audits later this year.

We have concluded that there are risks in relation to the processing of personal data by many political parties. Particular concerns include the purchasing of marketing lists and lifestyle information from data brokers without sufficient due diligence, a lack of fair processing and the use of third party data analytics companies, with insufficient checks around consent.

The ICO report was sent to the DCMS select committee which used it as part of its evidence. The committee issued a report, “Disinformation and ‘fake news’”, dated Feb 2019.  Damien Collins, the committee’s then Chair said on the landing page,

“Democracy is at risk from the malicious and relentless targeting of citizens with disinformation and personalised ‘dark adverts’ from unidentifiable sources, delivered through the major social media platforms we use everyday. Much of this is directed from agencies working in foreign countries, including Russia.

Damien Collins MP

Professor Dommett summarised the lecture as asking and answering the following questions. What is DCC? What do people do? Why do people do different things? Why are we concerned now?

She classifies political data as publicly available, personally disclosed, obtained from monitoring, and inferred. Her conclusion is that this is not scary, however I remain concerned about data obtained via monitoring and inference.

She states that party-political tech, outside the US, is poor, a conclusion backed up by “Systems Update Required”, which looks at the dreadful IT operation of the Labour Party for the 2019 general election. The US is a different place, which matters because many of the skills and software used for data-driven campaigning comes from the USA and they are not constrained by the GDPR. Actually, most software comes from the USA. She suggests that Trump was the originator of micro targeting; I would suggest that it was Obama’s first campaign where micro targeting was used for the first time, although anecdotally it would seem primarily for fund raising. She argues that the UK political parties do not micro target and that group targeting is a sophisticated as it gets.

The reasons for variations are based on differences in regulation, political strategy of the parties, and on the electoral system and structure. Obviously in the UK far more effort is applied to voters in marginal seats, and as a counter point to that, in Australia, for instance, due to compulsory voting, there is very little emphasis on the GOTV operations. It would be interesting to see if in Australia more work was done on persuasion.

The reason I look at persuasion and GOTV as separate is that I have come to the conclusion, that political campaigning breaks into these two functions. It is an argument conducted in the Labour Party in many constituencies, amongst its activists, about good campaigning practice. It also would seem that Labour, at least, has limited interest in persuading people to change their mind; their strategy would seem to be based on giving people what they want or at least giving marginal seat voters what Labour thinks they want. I have looked at what academia says about the effectiveness of doorstep work versus other B2C techniques and consider if its worth it. My conclusion is that while your opponents do it, you have to. Labour’s volunteer driven doorstep operation counters the costly direct mails of the Conservatives.

The second speaker was Miriam Sorace, who works as a researcher/teacher at the University of Kent and is currently working for the Labour Party. She grew up in Slovenia. I mentioned this, firstly because she did, but also during my work on the report on the Conference on the Future of Europe (CoFoE) I’ve come to the obvious conclusion the cultural focus of political speech and campaigning in Eastern Europe has been formed by the 45 years of dictatorship. I’m of the view that this experience needs to be heard, although they’re not necessarily right on all issues. While there is strong resistance to the state control of the media, there is little recognition that private monopoly ownership is equally unacceptable.

She also presented her central belief that campaigning needs to reach the undecided and is primarily about mobilisation aka GOTV. She claims that DCC is about mobilisation because, “humans are confirmation bias machines”. This means that the fear of fake information manipulating people is overrated since people find some lies either hard or easy to believe. i.e. it’s the manipulation that doesn’t occur. She also argues that prestige counts by which she means the records and reputations of those politicians making claims are important.

This takes us back to Labour’s doorstep campaigns with its Voter Id and GOTV operations. This has been complicated by the introduction of the “postal vote on demand” since the deadline is unclear. I am a big fan of understanding why what one does works; to me, this requires understanding some theory. I ask, and examine the evidence,  if doorstep campaigning is worth it in an article on my wiki, Is voter-id/gotv worthwhile? and appreciate the efforts of people like Dr Sorace to apply science and theory to problem solving. Erin Meyer identifies applications vs principles as one of the eight dimensions of business communication. i.e. what we do vs. why we do it. I have often been frustrated by the dismissal of well researched and scientifically proven proposals with the airy, “let’s use our common sense”. Meyer suggests that the British are neutral on this spectrum; that’s not my experience.

I felt this came close to, ethos, logos and pathos [or on Medium]. I misremembered it as including virtue as one of qualities that make up the pillars of rhetoric. What she said also reminded me of the “Need for enemies” (Fergusson, 2016) which suggests that politicians have little motivation for eradicating the problems for which they have a reputation of being able to solve. However, this all underestimates the secret empowerment of prejudice. i.e. you can’t reason people out of a position they haven’t reasoned themselves into.

The third speaker was Louise Edwards of the Electoral Commission

She started with some statistics: about £50 million was spent on the last general election, there were 400 parties and 130 campaigning organisations. Many of the parties were hyper local. Since Labour spent about £8½m, that’s an awful lot of money being spent by others. The Tories spent £16m.

Most complaints were about disinformation.

Campaign is generally good and so is the public trust. I don’t know how she can say this after the referendum remembering that the High Court ruled the crimes committed by the leave side would have voided the election if it has been a statutory mandate and the changes made by the Tories to election law.  It is clear to me, from casual study of the press, that the Electoral Commission do not have the powers, particularly the powers of sanction, required to deter criminality.

The EC spokesman said they were only interested in money, that foreign influence was the remit of the security and intelligence services, and I was reminded that the ONS can be used to fact check government ministers.

Louise shared that in the Netherlands there is a code of conduct for elections which unlike the UK requires politicians to tell the truth. She also reminded us that the UK’s interpretation of freedom of speech permits politicians to lie, although not about their opponents it would seem.

GDPR, not only requires accuracy and inferred data cannot be determined if accurate or not, it requires that people are not profiled by machines and political views are ‘special data’.

In conclusion, inadequate regulation is a problem and needs to be addressed. The opaque operations of the political parties is a problem, as is the question of foreign funding.

To me this is part of the problem, there is no one regulator, the EC has no effective power of punishment, the ICO is not interested in regulating the Parties, and the ONS can only correct  government ministers and does not have a rapid rebuttal capability and was shown to be ineffective during the referendum.


Here are some informal footnotes, some of which come as responses to the Q&A,

The panel was exclusively female.

The meeting chair challenged Miriam Sorace on the question of the confirmation bias, my notes say they hope that her statement is not true as it excludes or diminishes the need for persuasion based on truth.

Dommett emphasised the impact of non-voters, i.e. abstainers, either conscious or not, and the existence of safe seats within the UK. I note that marginal seat campaigning is no good in mayoral elections because people learn that their vote doesn’t count and fail to acknowledge when systems and boundaries change. Political parties’ focus on marginal seats or more accurately their ignoring of safe seats, their own and their opponents, legitimises apathy.

It was suggested that Gen AI might be the technology that enables microtargeting as one of the key constraints to doing this is the speed of developing content. Gen AI would help do this.

I got to ask a question, it was something like this, “The GDPR requires information to be lawfully obtained and accurate, how can inference attacks and/or the combination of lawfully and fairly obtained information with the information purchased from third parties be guaranteed to be accurate and lawfully obtained? Is this not an area where more regulation is required? (In retrospect the law may be good enough and it comes back to the ICO’s lack of will.) I sneak in a second question on the lack of transparency in particular how Facebook is used, I quoted the Copeland by-election and I wondered what the Tories had been saying about the Labour candidate and their commitment to nuclear power given the leader of the Labour Party was alleged to be an opponent of it. I also had a pop at the belief that people weren’t persuadable; they clearly are unless the falling Tory popularity is based exclusively on demographics. It’s a deeply pessimistic view, if true there is no point talking about policy and this takes me back to where I opened,  “other people have good ideas”.

Professor Dommett’s comments on effective technology reminded me of the document, “Systems update required” commissioned by Labour Together, to which I refer and link above. I blogged about it, in an article called “What does Systems Update required say about Labour’s IT?” [or on Medium]. In order to complete this part of the article, I returned to my article and extracted this quote from the it,

Labour no longer has leadership in the doorstep/campaigning apps and these, it would seem, are no longer considered competitive advantage; many progressive parties use either open source or third party programs to do this work. (I argue elsewhere and frequently, …, that one should only build software where the organisation’s functionality is unique and gives competitive advantage).

Dave Levy CITP, CIPP(E)

The most economic statement of my views on buy vs build is in my linkedin article, entitled, “Software Programme Management” [or a précis/pointer on Medium]. In it, I say, repeating and supporting Dan Remenyi,

Another insight is that it one has to understand if the business process is unique or commodity, …. A unique business process is likely to require unique software, where functions deemed ‘factory’ or ‘support’ may be best suited to be delivered by off the shelf software. An additional dimension of Remenyi’s thinking is that while, if one requires unique function, software should be developed, if not, it should be bought and that the optimum process as defined by the software author should be adopted.

Dave Levy CITP, CIPP(E)

Dave Levy, CITP, CIPP(E)

I remind you of my post nominals, to establish my authority and expertise on the privacy policy and technology parts of this article. See also,

  1. https://www.linkedin.com/in/davelevy/
  2. “Need for enemies” (Fergusson, 2016) @davelevy.info [or on Medium]
  3. What does “System update required” say about Labour’s IT @davelevy.info [or on Medium]
  4. Is voter id/gotv worthwhile @davelevy.info/wiki
  5. Meyers cultural map @davelevy.info/wiki  [or on Medium]
Data-driven campaigning: how and why do political parties do it?
Tagged on:                     

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: