The online-only Independent’s “front page” today breathlessly reported that “illegal Facebook spending ‘won 2016 vote for Leave'”. Philip Howard, a professor at the “Oxford Internet Institute” said it was his “professional opinion” that it was “very likely” that “excessive spending by Vote Leave altered the result of the referendum”. Howard is due to put his “evidence” to the High Court tomorrow…

On closer inspection, it appears “very likely” that Howard’s evidence is a load of utter rubbish, with statisticians having a field day tearing apart his numbers on Twitter. Howard’s calculations get off to a good start with his assumption that 80 million people saw Vote Leave Facebook ads during the so-called “period of excess spending”. That’s in a country of 66 million people, with an electorate of 46.5 million. By Howard’s calculation, 172% of the electorate saw Vote Leave’s ads…

Of those 80 million people, Howard then claims to use a “conservative industry estimate” that 10% of people who saw the ads would have clicked on them. Of that 10%, he claims that 10% believe, and then “a further 10% of that number can be expected to do something”. Except in his own calculation, he entirely emits his own final step, claiming that 800,000 people (10% of 10% of his imaginary electorate of 80 million) had their votes swung by the ads in the final week of the referendum, rather than 80,000 as his own model predicts. Leave won by over 1.2 million votes…

Moreover, the source he cites for his “conservative industry estimate” estimate of 10% click-through rate – a book called New Media Campaigns and the Managed Citizen – says “banner ads on politics topics generally had a 1 percent click-through rate”. The author of this venerable source? A certain Professor Philip Howard…

Even glossing over Howard’s other major errors, including… deep breath… conflating organic Vote Leave post impressions with paid ad impressions, assuming that the entirety of his imaginary electorate uses Facebook, ignoring the fact that Stronger In were also running huge numbers of Facebook ads and that more people made up their minds to vote Remain than Leave in the final week, assuming that everyone who decided to vote Leave after seeing an ad had previously been planning to vote Remain, the maximum number of people whose votes could have been swung – according to his own methodology – is 4,650. There’s wishful thinking, and then there’s pure and utter fantasy…

UPDATE: For any readers looking to enjoy more of Howard’s stellar “research”, Guido can heartily recommend this paper where he defines all Twitter accounts tweeting more than 50 times a day as bots. Although if you do fall into that category yourself, Guido might suggest getting a few other hobbies…