From Elsewhere: Economist magazine fisks biased academic report on the public reaction to the Jo Cox murder

The underpants of Dr Imran Awan and Dr Irene Zempi collectively combust following the revelation by the Economist magazine that their latest report is utter guff


Recently the Tell Mama group heavily promoted on their social media outlets a story from the Guardian that was heavily based on an academic report co-written by a member of Tell Mama’s advisory board Dr Imran Awan. The report alleged that there were tens of thousands of Tweets celebrating the murder of the Labour MP Jo Cox. The truth, as writers from the Economist magazine discovered, is very different from what these academics claimed was the truth.

Unfortunately for The Guardian and many others who took seriously the research by Dr Awan of Birmingham City University and Dr Irene Zempi of Nottingham Trent University, the claims of these academics have been pretty well busted by The Economist magazine. The report which was commissioned and therefore probably paid for by the Left-wing ‘Hope Not Hate’ organisation, stated in a press release about the report by Dr’s Awan and Zempi, that: “In the month after the killing at least 25,000 people sent more than 50,000 tweets celebrating her death or praising her murderer, Thomas Mair”. The Press jumped on the press release and the referenced report with the Guardian going very big on this story.” Unfortunately for the academics involved, The Guardian and others who uncritically ran with this story, things as the Economist discovered, were not all they seemed when it came to the data used in this report. To the Guardian’s credit they did remove the story when they realised it was false but other newspapers namely the Independent have not done the same and are still running this story as if it was true.

HOPE NOT HATE, a charity that combats extremism, published a report on November 28th purportedly revealing a mass outbreak of online hate-speech after the murder of Jo Cox, a Labour MP, a week before the Brexit referendum in June. In the month after the killing at least 25,000 people sent more than 50,000 tweets celebrating her death or praising her murderer, Thomas Mair, Hope Not Hate said.

Britain’s largest newspapers leapt to publish the shocking findings. The story was shared far and wide. Angela Eagle, a Labour MP, cast its conclusions as “staggering and appalling”. The news struck as Britain was coming to terms with an increase in hate crimes. (This week a man stabbed a passenger on a London train reportedly while shouting anti-Muslim slogans.) The discovery that such hateful attitudes were not isolated, and that tens of thousands of members of the public were willing to voice them openly, was deeply troubling.

Of course if it was indeed the case that there were 50k Tweets celebrating the murder of an MP then it would be a prime indicator that the country was in deep political trouble. If there were 50k people willing to go into the public eye and say ‘well done’ to Cox’s murderer Thomas Mair, then it would signify that a large number of people had lost complete faith in the Parliamentary system and preferred murder to ballot boxes. It would also indicate that if there are such a large enough number of people prepared to celebrate Cox’s death in public, then how many more must there be who were celebrating in private? It would be a worrying finding no matter what ones personal political view. But like anything to do with Tell Mama and it’s ‘tame academics’ like Irwan and Zempi, there is much less to this story than meets the eye. In fact the ‘narrative’ of the report, that there are a large number of violent ‘far rightists’, crumbles as soon as the writers at the Economist started to examine the data.

The Economist added:

An investigation by The Economist has found that Hope Not Hate misrepresented the findings of its own report when first releasing it to the press. The report itself gave a confusing impression of the number of tweets that celebrated Ms Cox’s murder. We estimate that, in reality, of hundreds of thousands of tweets mentioning the MP by name, the number that celebrated her death was at most 1,500, and probably much lower.

Fifteen hundred as opposed to fifty thousand is a hell of a difference between the number of celebratory Tweets that the Economist found and what the report’s authors claimed. It’s a third of what Irwan and Zempi are claiming.

The Economist continued:

Although press coverage of the story appeared to misread the report, that is not entirely the fault of journalists. The claim that 50,000 tweets celebrated Ms Cox’s death or praised her killer comes from the first paragraph of a press release sent out by Hope Not Hate ahead of the report’s publication. It does not appear in the study itself, which found only that a “majority” of the tweets, which related to both Ms Cox’s murder and the Brexit referendum more broadly, “related to specific calls for violence” (a term that is not defined).

So we have Hope Not Hate talking this up as well in their Press Releases. That doesn’t surprise me, Neither does it surprise me to find that the Guardian is taking Press Releases from Leftist activist groups like HNH and publishing them without any real process of fact checking taking place. The Guardian journalists should have taken the time to read at least an abstract of the report before publishing this questionable information. The majority of the Tweets did not celebrate Jo Cox’s death and the authors of the report appear to have just bundled in any Tweets that they considered relevant according their own prejudices and bias. According to the Economist Irwan and Zempi seemed to bung in anything that fitted the bill and then narrowed it down by searching for hashtags #refugeesnotwelcome and #DeportallMuslims.

Here’s the Economist on the methodology of the report. I think you may agree with me that Irwan and Zempi’s procedure seems slapdash and allows biases to easily creep in.

The Economist said:

In their study, entitled “Jo Cox ‘deserved to die’”, Mr Awan and Ms Zempi examine a sample of 53,000 tweets with hashtags related to both Ms Cox’s murder and the Brexit referendum. The total number of tweets on these subjects during the period was significantly more than 53,000; the authors appear to have selected their sample by narrowing their search using hashtags including #refugeesnotwelcome and #DeportallMuslims.

The report does not say what proportion of the 53,000 sample tweets related to Ms Cox’s murder, and what share concerned Brexit more generally. When The Economist asked the authors for help, they declined to share their data with us, citing death threats they said they had received since the report’s release. So we undertook our own analysis, examining tweets from June and July that included the terms “Jo Cox” or “#JoCox”—some 341,000 unique messages. Of a random sample of 800 of these, none was celebratory, and just four seemed to be derogatory toward Ms Cox, criticising her support for Syrian refugees, for instance. From this, simple statistics suggest that the true number of tweets cheering the politician’s murder would lie between 0 and 1,500. (The Hope Not Hate report reproduces about 30.) Mr Awan notes that our sample did not include tweets that mentioned only the killer, Mr Mair; it is also likely that some tweets were deleted before our collection.

The archive of tweets was gathered by Martin Goodson and Rafal Kwasny, data scientists at Evolution AI, a London-based startup. They found the Hope Not Hate report had other claims that seemed overdone. For example, it said a “key theme” on Twitter was the description of Mr Mair as a “hero”. In fact, many tweets containing the word “hero” were referring either to Ms Cox herself or to a pensioner who was injured while intervening to save her.

I must say that ‘death threats’ if the academics release their data for checking strikes me as very convenient for the Irwan and Zempi team. It means they never have to defend their views or disclose their data. What are these academics specifically concerned about that means they can’t email the data to the Economist or meet one of their writers? After all if you was trying to work out who was likely to be a ‘far right terrorist murderer’, writers from the Economist would probably be the very last suspects on your list. This bit of victim-hood claiming that may be false or not false, certainly makes me want to file Irwan and Zempi’s report into the round floor mounted filing cabinet. It’s guff and smells very strongly of guff. Irwan and Zempi along with HNH have put out a very dubious report and will not allow others to check to see just how damned dodgy it really is. That’s not the actions of trustworthy academics and the ‘death threat’ excuse is even more dodgy.

The Economist writer then went on to say how the Irwan and Zempi admitted that they’d been very broad with their categorisation and had lumped in a large number of tweets which may be described as ‘xenophobic’ with the much lesser number of Tweets celebrating Cox’s death. In other words Irwan and Zempi threw everything including the kitchen sink into the data pool and classified merely ‘xenophobic’ Tweets as being equivalent to celebrations of murder. This is a similar modus operandi as was (and probably still is) practised by the Tell Mama organisation following the Islamic murder of Lee Rigby. Then, Tell Mama counted people saying disagreeable things about Islam as being ‘attacks’. This allowed them to give the false impression that there had been a ‘massive rise in Islamophobia’ and actual ‘attacks upon Muslims’ following the murder. These false claims by Tell Mama were later exposed and contributed to the growing disdain and distrust that Tell Mama is now held. It seems that academics like Irwan and Zempi who either work with Tell Mama or are often quoted by them should have their work placed under the same level of suspicion as Tell Mama’s output is itself held.

It’s difficult to find a different word to describe this sort of biased report as anything other than dishonest. If the findings of the Economist are true, and I have no good reason to doubt that, then this report is so useless as a guide to online behaviour that it should be treated as at best influenced by conscious bias or at worst fraudulent. The Irwan and Zempi team appear to have gone out of their way to get the answer that they wanted rather than the true answer that the correct and relevant data would have given them. That is not to my view of an example of honest academic endeavour and is merely the obvious political hack-work that the Economist has so plainly shown this report to be.


Economist report into the ‘Jo Cox murder celebration tweet’ story

Tell Mama advisory board

The Guardian to their credit later realised that the information supplied in the HNH report was guff and removed the story

The Independent on the other hand is still treating this fake story as true