Why Carole Cadwalladr is a Conspiracy Theorist

Carole Cadwalladr has recently gained a high profile through a TED talk or two. She's been around for a while, digging up information on the activities of data analytics companies, especially Cambridge Analytica.

Some of my friends have quoted her approvingly on social media. I was less impressed. Eventually someone called on me to do a proper analysis of one of her articles.

Here it is. The article can be found at: https://www.theguardian.com/technology/2017/may/07/the-great-british-brexit-robbery-hijacked-democracy?fbclid=IwAR0FD3P6OeJDT1ObEyRPx8KHQUv-PK0LM6oRU9mdEgo8fj57e7Z7YdjhFj4

If you don't want to read the whole thing, my conclusions are in the last 4 paragraphs. (Unlike Carole who likes to put her conclusions at the beginning and then makes us wait ages for any justification!)

The article says at the beginning that Cambridge Analytica won the election for Donald Trump. The statement has no supporting evidence; however it is in a quote from someone (meaning it can be disowned if libel proceedings are raised). 

The same employee then claims that CA interfered in elections around the world in ways that would be illegal in the West. Again, no supporting evidence.

Then another firm gets mentioned - Palantir. No actual link is made between Palantir and CA but the language used implies guilt by association. 

This is followed by a statement that is obviously true (probably to make it look like the whole story is true) but that on its own has little impact ("a small handful of Silicon Valley firms are at the centre of the global tectonic shift we are currently witnessing.")

Then it makes another unsubstantiated statement ("Brexit and Trump are intertwined") and makes it seem believable by saying the same thing in several different ways. 

So much for the dramatic-sounding but potentially misleading introduction. Let's get down to the main story - or rather, the "three strands" of the story. Once again, each is presented as a fait accompli with no supporting evidence. One would hope that such evidence might eventually appear somewhere in the article.

First of all, she jumps to a claim that doesn't directly support any of her three strands: "Cambridge Analytica is a central point in the Right's 'propaganda machine'" (once again, presented as a quote from someone whose reliability isn't established beyond the fact he's a professor). She doesn't define 'central point'. She misleadingly mentions the Right immediately after describing some of the activities of the 'alt-right', thus implying the two are the same. 

She does, however, define 'propaganda machine' -- but only after linking two names to it, one being a high-profile Republican political figure and both being directors of CA, on the basis that she "believed" these two guys were linked to the "propaganda machine". That's not good evidence.

Then she makes her connection to Brexit - on the basis of a quote from the communications director of Leave.EU that Cambridge Analytica was directed to "help" the Leave campaign. That quote is actually the highest quality evidence she has presented for anything so far. 

After patting herself on the back for writing an article that triggered investigations into Leave.EU, she makes her most specific allegation so far -- that four Leave-preferring campaign groups spent money with one Canadian data company and that this constitutes collusion. She then states that 'Vote Leave says the Electoral Commission “looked into this” and gave it “a clean bill of health”'... in other words, there is evidence that her accusation of collusion has been investigated and found to be false, but she wants us to disbelieve that so she puts the statement into the mouths of Vote Leave and puts the Commissions's investigation and findings in inverted commas. Much later in the article she expands on her scepticism by quoting 'sources close to the investigation' -- so not even attributable quotes. 

Next she makes the most meaningful statement in her whole article to date: that Britain's electoral spending laws are no longer fit for purpose. It's a statement that is pretty clearly true but doesn't provide any evidence towards her various unsupported conclusions. 

Then she alleges that CA and the Canadian data company were intimately entwined. Once again this is presented as a quote from a former employee. She presents no further evidence beyond that.Next she heads down yet another side branch, claiming that CA's parent company SCL was "part of the British defence establishment" because it has a "Director of Defence Operations" who was a former military head of psyops. That's waaay too strong a claim. Lots of companies contract to the Ministry of Defence and lots hire ex-military personnel to help them win and manage contracts. The most she has actually established is that SCL made plans to win contracts from the MOD and that they had a psyops expert on board. 

Back to the main story: she claims that CA collected loads of data from personality quizzes on Facebook and from other stories and used it to create profiles of individual voters, who could then be sent targeted messages. This is pretty much established as truth and has been decried since as unacceptable, causing CA to go out of business. (And it's now specifically banned by GDPR). 

She then (finally!) gives some evidence for one of her 'three strands': "How British democracy was subverted through a covert, far-reaching plan of coordination enabled by a US billionaire." Actually, she isn't talking about British democracy yet, and all she has established is that CA had the data to send people highly targeted messages, which doesn't prove that democracy was thereby subverted. But she does manage to get the billionaire's involvement in. 

She does try to prove her point about subversion (which is critical to her whole argument) with yet another quote from someone -- an associate professor of philosophy (!!!) I would have expected a full professor of psychology at least for such an important point. Maybe she thinks that by the time people have got this far through the article they will have given up checking whether her quotes come from reliable-seeming sources. 

Then she's off on a side channel again, this time claiming that CA and the Canadian company supported the development of a Minority Report-like system in Trinidad in 2013. Apart from smearing CA's moral standards, the only thing this proves as far as the main argument is concerned is that CA and the Canadian company had had a close working relationship. So she's generated some evidence to support one of her other side branches. 

Now she jumps from this side channel to one of her three strands -- that the US is laying a basis for an authoritarian surveillance state. Her argument is that the same company that did the thing in Trinidad (which she calls 'the company that got Trump elected', a claim she still hasn't proved but seems to be trying to establish by repetition) has been awarded contracts by two US government departments - Defense and State, Oh, and the political guy who was director of CA is now high up in the White House. This is nowhere near enough evidence that the US is laying a basis for an authoritarian surveillance state -- if one of my students tried to make this link, I would fail them. If we had a clue what the contracts were for or how big they were, that might give us a clue -- but in the US as in the UK, all sorts of companies work with Defense. We can't assume that building data profiles of individuals suitable for 'Minority Report'-style monitoring was the only kind of work that CA did, whereas the author apparently does assume that here.

In her next paragraph she says "Documents detail Cambridge Analytica is involved with many other right-leaning billionaires, including Rupert Murdoch". What documents? Involved how? She gives one example -- CA tried to place an article in the Murdoch-owned Wall Street Journal and Murdoch himself allegedly had some involvement. So what? 

Now she drags Palantir back into the story -- because they too are owned by an (allegedly) right-leaning billionaire. And they too do data analytics. I think this must be her evidence for her third strand -- that "we are in the midst of a massive land grab for power by billionaires via our data." It's weak evidence, to put it kindly. 

Then it's back to Vote Leave and the Canadian company. Her primary allegation is that the link was forged because Vote Leave's Chief Technology Officer was a former CA employee. This is one of her better pieces of evidence.

Next is a graphic showing how everyone is allegedly linked. It's misleading (again) -- firstly for including Donald Trump (who hasn't been mentioned as having any role so far); secondly by linking Trump directly to CA as well as through Steve Bannon (there's no evidence AT ALL for a direct link in the article); and thirdly by including a firm and a person who haven't previously been named (another thing my students get marked down for). 

I'm going to stop my analysis here, except for one thing. Apparently the £100,000 sent to the Canadian firm by Veterans for Britain led to "a small number of people they identified as “persuadable” [being] bombarded with more than a billion ads, the vast majority in the last few days."

Seriously??? A 'small' number of people -- let's be generous and say 'small' is 10,000 people. They have been sent more than a billion ads? That's 10,000 ads each! No way could even the most dedicated social media user have seen 10,000 ads in a few days, never mind Veterans for Britain. Even if it's more than 10,00 people, or the article is using an American billion rather than a British one, we're still talking hundreds and hundreds of ads in a few days. 

My conclusion: This is indeed a conspiracy theory because it alleges covert collusion between multiple individuals to do something/things that was at best unethical and at worst illegal. 

The article has three 'strands'. None of those strands come close to being proved by the evidence presented. One is based on a lot of supposition about the nature and extent of CA's contracts with Defense and State in the US. One is based on the views of a mid-level academic in philosophy about psychology. And the third is generalised from two examples -- and it's not even clear that the allegation is true in one of those two cases.

And then there's the fourth claim -- that Cambridge Analytica got Donald Trump elected. Repeating the claim several times without presenting any supporting evidence does not make it true. 

Finally the article is misleading on several occasions. I cannot prove that it is deliberate but I would be very, very surprised if it isn't.