Michael Nugent Tackles @Demos’ ‘Islamophobia’ Report


Last week, I reported on the flawed research into online ‘Islamophobia’ carried out by the think tank ‘Demos’. The findings of this report painted an alarming picture of hate towards Muslims on the social networking site Twitter. The Demos report was spread far and wide unquestioningly by countless media outlets including the BBC.

I highlighted an excellent response from Benjamin Jones of the National Secular Society which detailed the unhelpful nature of the term ‘Islamophobia’ and Demos’ inconsistent criteria for classifying it.

Much to his credit, Carl Miller of Demos then responded to these criticisms in detail over on his website. Whilst I commend Carl for his willingness to engage openly on this issue in a constructive, civil manner, I also felt that he didn’t really address the key concerns.

As has come to be expected of Michael Nugent of Atheist Ireland, he has produced an incredibly diligent and comprehensive piece of reporting which further highlights the problems with Demos’ research and Carl’s response to the criticism of it. Please read it in a full here, but I will highlight some key points:

On issues with the Term ‘Islamophobia’:

Carl from Demos misunderstands the main concern about the use of the word ‘Islamophobia’. He comes closest when he says that: “The crux it comes down to is this. Many of the critics, I think, just don’t like Islamophobia as a concept at all… In my point of view, Islamophobia is real. It exists, it is a problem, it is not a crime, and it’s well worth researching.” That’s nearly right, but it conflates the concept and the description of the concept.

We of course know that anti-Muslim prejudice is real, it exists, it is reprehensible, and it should be researched and challenged. But we reject the use of the word ‘Islamophobia’ to describe that concept, because the Organisation of Islamic Cooperation and others also use that word as a propaganda tool, in order to silence legitimate criticism of Islam as an ideology, and thereby to buttress the human rights abuses that Islamist regimes impose on Muslims and non-Muslims alike.

Also, you can more effectively research and challenge anti-Muslim prejudice by describing it (far more clearly, accurately and unambiguously) as ‘anti-Muslim prejudice’, thus avoiding the conflations and harm caused by using the word ‘Islamophobia’. So why not simply do this? Why use a description that you know will cause complications and confusion, when you can use a clear description that you know will not do this?

On the criteria for ‘Islamophobia’, premature trends and the Texas Sharp Shooter fallacy:

Of the fifty keywords Demos used to collect “tweets that could be derogatory and anti-Islamic,” four are words commonly used in reasonable discourse, two are derogatory words used by some Muslims to describe non-Muslims, three are names of anti-Islamic political groups, one word is abusive of Islam, eight are examples of anti-Muslim bigotry, and thirty two (almost two thirds) are racist or xenophobic rather than religious abuse.

This leads the report to conflate as ‘Islamophobic’ tweets that are polar opposites to each other. It gives as two examples from 5 July: “Morocco deletes a whole section of the Koran from school curriculum as it’s full of jihad incitement and violence. The Religion of peace”; and “I fucking hate Pakis.” The first tweet here is a reasonable and ethically good analysis of a harmful ideology. The second is reprehensible hate speech against people based on their ethnicity. As Sesame Street used to tell me as a child, one of these things is not like the other.

Even using its own preferred definitions, and its own data of 34,000,000 tweets, the report could have chosen to describe its key finding more constructively as: “Even in politically provocative times, up to 99% of discussion of jihad, terrorism and hijabs on Twitter does not demonstrate prejudicial dislike of Muslims because of their faith.” This would have been a significant, and accurate, lead finding. Instead it is relegated to an implication of the Methodology section on page 18 of a 21 page report.

Instead, using the less than 2% of their original data that they concluded were actually ‘Islamophobic’, one of the key findings is melodramatically phrased as: “Islamophobia on Twitter is increasing month on month – with July the highest rate since Demos’ dedicated analysis began in March.” But they are talking about 19 pilot days in March and April, and only three months full analysis, with monthly fluctuations of minus 50% to plus 27%.

That’s not enough data to establish any trends, let alone lead with them as a key finding.
Also, the report does not highlight that we are talking about a percentage of a percentage of the overall daily traffic of 500 million tweets, and that we thus risking giving undue prominence to abusive bigots whose conversational repertoire includes ‘camel fucker’, ‘dune nigger’ and ‘sand monkey’.

Of course we should highlight this reprehensible behaviour in order to challenge it. But why not also highlight, and indeed lead with, the finding that the vast, vast, vast majority of Twitter discussion of even controversial topics like jihad, terrorism and hijabs does not display anti-Muslim prejudice? As well as being accurate, it would help to dampen down tensions rather than inflame them.

Demos’ own Inconsistent definition of ‘Islamophobia’:

…the report’s definition is subjective and ad-hoc, as indeed are all uses of the term ‘Islamophobia’. It does not match with other definitions of ‘Islamophobia’, including those used by sources that the report’s authors cite in defending the report, or those used by the Organisation of Islamic Cooperation who also use the term to silence legitimate criticism of Islam. As usual, the term ‘Islamophobia’ is a moving target of whatever you are having yourself. As usual, when a well-intentioned group like Demos uses the term to mean one thing, other groups will use their report to justify and buttress something entirely different and more sinister.

…the report’s definition does not appear until the Methodology section on page 16 of a 21-page report. The body of the report instead casually interchanges three very different terms: “anti-Islamic activity” (whatever that might be), “anti-Islamic expressions” (which may be both reasonable and ethically necessary, depending on their context) and “anti-Islamic hatred” (which on the face of it is hatred of the ideology of Islam as opposed to dislike of Muslims because of their faith). The report then collectively describes these three very different phenomena as ‘Islamophobia’. So, even within the same report, the term ‘Islamophobia’ is a moving target of whatever you are having yourself.

There are many more illuminating findings in Nugent’s excellent report. I urge you to read it in full here and share it far and wide. I look forward to Demos’ response. You can listen to my interview with Michael Nugent below:


Stephen Knight is host of The #GSPodcast. You can listen to The Godless Spellchecker Podcast here, and support it by becoming a patron here.


  • Islamophobia: A word, created by fascists, and used by cowards, to manipulate morons.
    In countries where Islam is the state religion or is adhered to by the majority of it’s population, the term Islamophobia is non existent. Why? Because it need not exist. Blasphemy laws take care of that. Criticize Islam and you risk government mandated fines, imprisonment, torture or death.
    In a country like Canada or the US, since it’s not legal to imprison or execute people who criticize religion, you can do the next best thing: Call them an Islamophobe. Because as everyone knows, if you criticize a violent, hateful, dangerous religion like Islam, it obviously means that you hate and/or want to kill every single last Muslim on the planet.
    Then when you point to examples of how hateful this religion can make people behave – a la Anjem Choudry, you’re instantly told “that’s not real Islam” or “what he says has nothing to do with Islam” or some other such idiotic garbage.
    The whole thing is designed to completely stamp out discussion of any kind and it’s working extremely well, because we have entire sections of the population, who are not Muslims themselves and know nothing of Islam, yet go out of their way to make excuses for the worst examples of human behavior that can be directly attributed to this evil religion. In their rush to appear tolerant, they put themselves in bed with the most vile and intolerant human beings on the face of the earth.
    Instead of Demos commissioning a report about “hateful tweets against Muslims” or some other such self-pitying bulls*&t, why not commission a report on the demographics of people who “honor” kill members of their own family? Or perhaps a study of people who kill because they felt their religion has been insulted (i.e. blasphemy killings). Let’s see what conclusions we can draw from studies like that shall we?

  • I assume being on Victoria Live, he could get away without having to go anywhere near an explanation of the word etc.
    If he was on Panorama, I suspect he’d be grilled about it.
    And perhaps Catrin Nye needs a bit of help with this. Her statement after the interview was interrupted, “…an interview on Islamophobia, is interrupted by…..Islamophobia,” has a definition of that word, nowhere near what Miller eventually came up with.

  • Statistics and damn lies…I suppose its part of the political world we have to live with but sadly there are not enough people like you who take the trouble to expose the bias behind them. I wonder how many of those ‘hate’ tweets came from the same person?

What do you think? Leave some comments!