Tom Chivers

Tom Chivers is a science writer. His first book, The AI Does Not Hate You is out now.


A year ago, the man who is now Prime Minister was describing Muslim women in burqas as looking like “letterboxes” and “bank robbers”. It was in a column for the Daily Telegraph1. Inevitably enough, it caused an uproar, because it was a stupid and ugly thing to say.

The strange thing is that the piece was a defence of Muslim women’s right to wear the full veil, and a condemnation of the Danish government’s decision to ban them. It was a couple of nasty little throwaway lines that caused the outrage, rather than the content itself.

Now, the NGO Tell MAMA, which records anti-Muslim incidents, has said that the piece was followed by a 375% surge in abuse and attacks on Muslims. The Guardian goes further, saying that the comments “led to a surge in anti-Muslim attacks”.

That sounded absolutely extraordinary to me. A 375% increase is enormous. So I thought I’d have a closer look.

The report records all “anti-Muslim incidents” received by Tell MAMA. That includes offline and online incidents, and also both verbal abuse and physical assaults. Tell MAMA says that it received 1,282 reports of anti-Muslim incidents in 2018, and that it “verified” 1,072 of them, 745 of which were offline.

Johnson’s “letterboxes” column was published on Monday 5 August 2018. In the week before that, Tell MAMA “received 8 reports of street-based (offline) anti-Muslim attacks”, by which I think they mean all incidents. But in the week afterward, the 6th to the 13th, “this jumped to 38, an increase of 375%”.

If I’ve read it correctly, and they’re referring to all verified offline incidents, then 745 incidents a year translates to a bit more than 14 a week, on average. So the eight reported the week before was anomalously low, making that 375% increase sound much more dramatic.

Still, 38 is 165% above the average. But, again, I can’t work out whether that’s unusual. The report doesn’t give weekly breakdowns, so I can’t see whether the numbers are usually pretty steady around the average, or if they jump up and down a lot. If the former, then a week of 38 is surprising; if the latter, it might not be. (I have asked Tell MAMA for more information, but so far, I haven’t heard anything; I’ll add an update here if I hear back from them.)

They do have a breakdown by month, which shows in August there were around 70 of these incidents, about the same as April, May and June, and rather lower than March, which had a little over 80. The Tell MAMA report says that the March spike was driven by a horrible letter-writing campaign called Punish a Muslim Day which called for anti-Muslim attacks on the 3 April.

I think, though, that the really standout thing about these numbers is how tiny they are. For instance, there were only 327 verified anti-Muslim incidents online in the whole of 2018. I can’t believe for a second that there aren’t more than that every day on Twitter alone, if “anti-Muslim incident” includes abusive language.

There are more than three million Muslims in the UK. Tell MAMA recorded about 1,000 anti-Muslim incidents, on and offline, and said that the police recorded nearly 2,000 more (they say there’s no overlap). If those numbers are the whole story – and Tell MAMA are very clear in saying that they’re not – then at most one in 1,000 British Muslims were the victims of anti-Muslim abuse of some form last year. I would be amazed if that’s not a huge underestimate.

What that means is that it’s much less likely that the Tell MAMA numbers are recording real underlying change. It’s like an opinion poll – the smaller your sample, the less likely it is to be representative of the population. In real opinion polls, the pollsters try to get around that problem by finding a representative sample, and weighting it where it is not.

The Tell MAMA numbers come from people who contact Tell MAMA, which is a non-random, non-representative sample of the population. It could, for instance, be simply that, in the aftermath of high-profile events such as Johnson’s column, anti-racist projects such as Tell MAMA get more publicity and so get more reports. Given the difficulty of teasing out media effects on real-world behaviour, that sounds more likely to me than a single column, behind the Telegraph paywall, causing a 375% spike in incidents.

The question, of course, is what is really happening with anti-Muslim hate crimes and abuse. Unnervingly, hate crimes recorded by the police have gone up dramatically in the past few years: from a little over 40,000 in 2011/12 to nearly 100,000 in 2017/18. Racially motivated crimes have gone up from 35,000 to 70,000; religiously motivated ones from 1,600 to 8,300.

But those increases, according to the police themselves, are largely due to the police getting better at recording these sorts of crimes, and a growing public awareness of them leading to more reporting. A Home Office hate crime report, using the Crime Survey for England and Wales data, reports that the total number of hate crimes has in fact dropped, from around 300,000 in 2007 to about 184,000 in 2018, although religiously motivated incidents have stayed roughly steady at about 38,000.

Interestingly, there was some evidence (not from the CSEW but from Home Office figures) of some spikes after major incidents – but those incidents were huge national news stories, such as the EU referendum or the five terror attacks in 2017, not pay-walled columns. And even they aren’t anything like the 375% increase Tell MAMA claims; they’re larger spikes in a spiky line.

So the problem is both less and more than the Tell MAMA report suggests, I think. Boris Johnson’s reckless column gets far more notoriety than it deserves; it may have caused a few isolated incidents, but I would be astonished if it had the sort of major effect it claims. But hate crimes in general are much more common than the small numbers Tell MAMA records.

Tell MAMA aren’t a polling company: they exist to give Muslims a place to report the abuse and attacks they find themselves on the end of. I don’t blame them for trying to collate that data and turn it into something useful. But the media should know better than to take the dramatic numbers it comes up with at face value, assume a causal link where none can be shown, and then plaster the whole thing under a dramatic headline. Britain already feels like a powder keg, and these are just more sparks flying around.

FOOTNOTES
  1.  I notice that the piece in question refers to a Danish anarchist commune called Christiania, “where I remember spending a happy afternoon 25 years ago inhaling the sweet air of freedom”. Since Christiania is, like Nimbin in northern New South Wales, a place where drug laws are ignored and people sell weed on the streets, I can’t help but think that Johnson is suggesting that he may have got extremely stoned in Copenhagen some time around 1993. I’m a little surprised that that didn’t get more attention, since while he’s admitted taking drugs – both cocaine and cannabis – before and during university, as far as I know he’s never admitted to doing drugs since. And at the time, Johnson would have been the Brussels correspondent for the Daily Telegraph.