BBC on grooming gangs

[ArabellaScott]

Interviews with women who survived grooming and abuse. Content warning.

Jess Phillips confirmed yesterday the report on grooming gangs has been delayed.

A new NPCC report is mentioned, too. Will try to find it.

www.bbc.co.uk/news/articles/c62n72mj113o

I think the NPCC data was not shared punlically, just with the BBC.

'The BBC has obtained new police data on the ethnicity of suspects arrested on suspicion of gang grooming offences in England and Wales in 2024.

The National Police Chiefs' Council (NPPC) figures indicated that British Pakistanis were substantially over-represented among suspects.

In 2024, of grooming gang suspects where an ethnicity was recorded, just over half were white British, and around one in eight were British Pakistani - even though one in 40 people in England and Wales were of Pakistani heritage according to the 2021 census.

The NPCC told Newsnight that the figures should be regarded with caution because just under a third of suspects had their ethnicity recorded.'

Either they need to record ethnicity every time or they need to not record it at all, I'd say. Partial stats like that are no help to anyone.

NPPC 😂. What have pigeons done now?

More seriously, having worked with and for Gov’t Departments partial statistics are extremely useful for obscuring the true picture.

I suspect ( given my previous work.) that ethnicity would be less likely to be recorded if the alleged perpetrators were anything but White British Male.

In this instance, partial stats look just as likely to further inflame the situation, and cast no clear light on it, either.

Well there is nothing like a bit of sunlight and that was nothing like a bit of sunlight.

I'm wondering if this was a leak to test the response, before the government release the report proper.

It was supposed to be a rapid three month review, and is due to be a month late.

https://www.bbc.co.uk/news/articles/cx2qjd95q1eo

I am genuinely confused by your comment

“partial stats look just as likely to further inflame the situation.”

If your ignore my previous comment about where one of my previous roles involved providing/ collating statistics / reports for National/Local Government Projects/ Pilot Programmes, EU Funded Projects, Mayor of London Funded Projects and how I was “encouraged “ to provide the “correct “ answers.

Then purely from a statistical perspective if you have a sample size of approximately 1/3 of the population. In this case the population being the ethnicity of people involved in grooming gangs then you can with a very high degree of confidence extrapolate your results to accurately reflect the ethnicity of all grooming gangs.

If you compare the ethnicity of grooming gangs to the total population and there is a statistically significant difference then again you can with a certain degree of confidence draw assumptions from your results.

Certain people/ groups may not like the answers produced but that is not the purpose of the analysis.

For example, the Guardian Newspaper gives minimal coverage of the size and scale of the grooming scandal across the UK but covered the Scottish grooming gang in great detail. I wonder why🤔

Then purely from a statistical perspective if you have a sample size of approximately 1/3 of the population. In this case the population being the ethnicity of people involved in grooming gangs then you can with a very high degree of confidence extrapolate your results to accurately reflect the ethnicity of all grooming gangs.

Ah, sure. If one assumes that the data collection/lack of collection was entirely random. My assumption was, and what I'd inferred from the NPCC comment, that it was possible that data collection itself might have been profiled/biased.

If one only collects data when the accused are of one ethnicity, then it could skew the stats.

But I know absolutely nothing about the methodology of their data collection.

Exactly. We do not know if the other 69% were all Pakistani and it was specifically chosen not to record their ethnicity for that reason,

Unfortunately I did not articulate my thought processes clearly and for that I apologise.

My first post was based upon the assumption that the source data was selected to support a certain agenda.

My second post was based upon the idea that the source data was randomly selected.
I chose to assume a normal distribution rather than assume that the distribution was skewed ( which in reality it probably is)

The interesting point is that even if you assume scenario 1 the evidence still shows that the ethnicity of grooming gangs does not reflect the demographics of society as a whole.

Obviously, the complete lack of information on the methodology used makes it very difficult to confidently extrapolate any meaningful analysis from this report.

It's not helpful either way, is it?

If the source data was selected with bias, the information is wrong. If the source data was random, then it hints at a social problem that is not being addressed.

Either there is a problem being covered up, or a problem artificially created. Neither is a good thing.

It is an issue that is constantly raised with TRA presented data too - how the sample is selected. It is generally some sort of convenience sampling eg those accessing a LGBT+ self-help website, or snowball sampling where one participant passes on the survey to contacts who in term pass to their contacts. This methodology is clearly very biased. But even where you have randomised sampling there can be bias in who chooses to answer and what answer they give - this has historically been seen with ‘shy Tories’ underestimating the Conservative vote in polling data.

Given the attempts at covering up the issue of grooming gangs, including by those in authority within the police, data from those of a specific ethnicity not being collected on purpose is a very real possibility.

I can imagine it could be biased either way. Or even both at once. Some forces/individuals may be collecting data to prove that some demographics are over represented. Some forces/individuals may be collecting data to prove the inverse. Some may be avoiding the subject altogether, for various reasons.

It will only be useful - and I mean that in the most objective way possible, useful to society, to either address a problem or put to rest a claimed problem that is not in fact a problem - if we get accurate data that has been collected in as objective manner as possible. Which I imagine would mean always collecting the ethnicity info, in a standardised format.

Which I imagine would mean always collecting the ethnicity info, in a standardised format.
You'd hope that the responsible organisations would do that. Because have a little bit of data is almost worse than useless: it hints at a problem just enough to inflame reflexive racists, without providing enough data to allow a sensible and targeted response to the culprits, which will inflame both sensible people and reflexive racists.

Doubles all round...

https://www.leighday.co.uk/news/news/2024-news/police-have-failed-to-record-information-about-race-or-ethnicity-of-more-than-33-000-people-referred-to-prevent/

“Information about people’s race or ethnicity is sensitive data under the law, and sloppy handling of that data is illegal. The police are in the worst of both worlds: they are attaching racial or ethnic labels to people, which creates serious data privacy issues, while doing it so erratically that no one can assess whether Prevent is having a discriminatory impact.”
...
The Information Commissioner’s Office has previously recognised the significant public interest in disclosing information relating to Prevent ethnicity data. Our client is concerned about the manner in which public authorities are dealing with this data: there appears to be a pattern of obfuscation and inconsistency across different public authorities in relation to how this data is collected, held, tracked and shared. There also appear to be substantial obstacles to obtaining this data, despite the clear public importance in understanding it. There is an ongoing debate about the appropriateness of the Prevent programme and its potentially disproportionate impact on minority ethnic groups. Our client considers that the ability to review, analyse, and discuss the ethnicity data would make an important contribution to that debate.”

(my bold. Different social issue, same problems with data collection).

The government know who have been charged, prosecuted and found guilty. The guilty should be on the sex offender register, shouldnt they? Even if their ethnicity wasnt recorded at the time, its not difficult to find.

It wont be as comprehensive as it should be, but its a good number to get an indication of any patterns around perpetrators, victims and how various bodies dealt with the situation.

There is an ongoing debate about the appropriateness of the Prevent programme and its potentially disproportionate impact on minority ethnic groups.

What constitutes a disproportionate impact?

There was this study. I haven’t gone through it properly so critiques welcome, but it found 83% of those charged with ‘group localise child sexual exploitation’ between 1997 and 2017 had ‘Muslim names’.

https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3248665