THREAD (PART 3 of 4): Why you shouldn’t trust @QuilliamOrg ‘s report on #groominggangs & the claim that 84% of “grooming gang offenders” are Asian. An evidence-based critique. 1/
It ignores well-known biases in crime data, such as whether victims disclose (‘self-selection bias’), whether reports are investigated/prosecuted (‘institutional bias’) & whether convictions make the news (‘media bias’) 2/
If using news sources, it's a big mistake to overlook #mediabias – especially as coverage of “grooming gangs” has been very racialised. Crimes involving sex, children & race ‘sell' & stories fitting an established narrative tend to get more coverage. (bit.ly/2q7WLKq) 3/
Despite all its major flaws, the report asserts that “the margin of error resulting from these limitations is minimal and negligible.” I think there’s some confusion here about what “margin of error” means (not this). 4/
I assume the basic point being made is that the results are reliable. After all, the report does claim “conclusively irrefutable” overrepresentation. This claim is not justified. The results are simply not credible for the various reasons discussed here. 5/
It is deeply ironic that a report full of spurious claims warns specifically against “false assurances” that allow debate “to be hijacked by those who may wish to promote their own political agenda and malign entire communities based on the actions of a few.” 6/
Strangely, much of the report is just a crude rehashing of two @CEOPUK reports, particularly playing up a supposed distinction between “Type 1” & ”Type 2” offenders. The idea offenders neatly divide into these two “types” isn’t necessarily very helpful or convincing. 7/
Tellingly, @CEOPUK publications don’t seem to have used this distinction since. Research suggests that, like human behaviour in general, child sexual abuse probably results from *interactions* between inherent disposition & available opportunities (bit.ly/2PxXEu8) 8/
In the report, @QuilliamOrg also appears to introduce its own unsubstantiated claims, such as that “Type 2 offenders cannot be defined as ‘grooming gangs’ because there are no grooming tactics involved”. 9/
Bizarrely, @QuilliamOrg ‘s own “data” section features lots of @CEOPUK results. Yet they neglected to include @CEOPUK ‘s own important warnings that their results were not representative & based on limited & patchy data. 10/
Overall, @QuilliamOrg ‘s report misinterprets, misattributes & misrepresents @CEOPUK ‘s work. By drawing on it so heavily, are @QuilliamOrg trying to seek credibility by association & detract from challenges? 11/
The “analysis” section is little more than an exercise in speculation about factors in Muslim & Pakistani culture supposedly driving abuse in the UK: #childmarriage #FGM #homophobia #misogyny #womensrights #integration #politicalcorrectness & selective reading of the #Quran. 12/
These issues are real but I question how they are positioned as the logical explanation for the report's limited results. This whole section was poorly-grounded in the results & felt very sensationalist to me. 13/
Even if correlation had been established (& it hadn’t), we should be sceptical of simple “cultural” explanations. Culture is important but culture alone cannot explain why the majority does not commit such offences, or so many from other cultures also offend. 14/
Later in the report, ten “case studies” are included. The stories are utterly grim but the cases are certainly not “analysed in depth” as claimed. Instead this section is little more than a (highly selective) rehash of news reports. 15/
Once again, evidence is selected to support claims & relevant counter-evidence conveniently omitted, such as relevant judges’ comments that they did not believe offences were racially motivated. (e.g. ind.pn/2SQArCc ; ind.pn/2PNAFLS ) 16/
The idea that certain cultures shouldn’t be “out-of-bounds when it comes to criticism” seems to me little more than a straw man argument. Of course they shouldn’t be “out of bounds”, but neither should we pathologise them through gross generalisations. 17/
References to the “Politically Correct Brigade” & “regressive left” seem invoked to dissuade rational debate & detract from criticism. Much of the abuse I’ve had from @QuilliamOrg 's @MaajidNawaz since challenging the report’s credibility has been framed accordingly. 18/ 
Reading the report, I can well imagine how its vivid images of a #clashofcultures & its talk of “divisive, unevolved cultural identities” might enrage the likes of Stephen Yaxley-Lennon (“Tommy Robinson”) & @UKIP ‘s @Aalancraig . 19/
Yaxley-Lennon cited @QuilliamOrg ‘s report in his defence to contempt of court, apparently to justify racial stereotyping (bit.ly/2NYnra7). His acts risked causing brave victims additional trauma & affecting their chance of getting justice (ind.pn/2F3HlRS). 20/
Meanwhile, @AalanCraig cited the report when he infamously spoke of a “holocaust of…our daughters”, saying @QuilliamOrg ‘s work on grooming had “traced a major part of the influence back to the Islamic faith” (bit.ly/2F7Ht2w) . 21/
To that, @MunaZainab reportedly responded “I do not condone this alarmist language & we did not use such language in our report”. While the abuses covered in the report *are* horrific, @QuilliamOrg 's repeated use of terms like “crisis” & “epidemic” seems to me alarmist. 22/
This point is especially true when you remember they reported on just 264 convicted offenders over 12 years & across the whole UK. Yes, the crimes were absolutely awful but we must see them amid the bigger picture. Surely the true epidemic here is *all child sexual abuse*? 23/
In 2016/2017, the police in England & Wales recorded over 13,000 rapes of children & several tens of thousands of other child sex offences (bit.ly/2OiOMUj). Only a fraction of child sexual abuse is reported: studies suggest as low as 3-15% (bit.ly/2SE9VvR). 24/
Court data show Asians weren’t overrepresented among roughly 172,000 men & 27,000 women convicted of sexual offences in England & Wales in 2016 (bit.ly/2H2dLfa). 25/ (Continued in fourth & final thread)
Here are links to the rest of the threads:
Part 4:
(And back to
Part 1:
Part 2: )
Part 4:
(And back to
Part 1:
Part 2: )
