A caveat with all data, but most acute in the case of suspects (many/most unnamed), is the difficulty of identifying repeat offending and therefore distinguishing the numbers of suspects (incl repeats) from the numbers of people (excl repeats).
A couple of details to add to this: Gwent, Durham and Suffolk have conspicuously low CDR rates for Black people. Across the 3 years, Gwent Police S&Sd 345 Black people (one every 3 days), Durham 63 (one every 17 days), and Suffolk 727 (one every 1.5 days). The MPS S&Sd 143 Black people/day on avg.
The details in this thread may also be of interest bsky.app/profile/gmha...
Counterpoint: they tell us a lot about what is reported by the public (and if you dig into the details, there are further wide differences in terms of things like suspect age profiles by offence type.)
Evaluation of a pilot involving structured supervisor reviews of officers' stop and search body worn video recordings - now being rolled out across the MPS www.met.police.uk/police-force...
Just checking you've seen the pilot evaluation report? www.met.police.uk/police-force...
I'll make some enquiries
Notably, "arrests data highlight the way the impact of non-borough visitors varies widely by borough and crime type (being highest in Westminster and for theft person), with important implications for the validity of disproportionality calculated [based on] the demographics of borough residents."
@stopwatchuk.bsky.social
@lisatompson.bsky.social this thread may be of interest
This is relevant, from the papers for today's London Policing Board session bsky.app/profile/gmha...
Apologies, I reposted my thread bsky.app/profile/gmha...
Notably, "arrests data highlight the way the impact of non-borough visitors varies widely by borough and crime type (being highest in Westminster and for theft person), with important implications for the validity of disproportionality calculated [based on] the demographics of borough residents."
This section may be of particular interest in the context of today's new report from MOPAC on disproportionality:
More on disproportionality from the S&S Charter report
From the Stop and Search Charter report at item 8, details of how the Met will be using structured supervisor reviews of body worn video footage to "track the quality of searches at an organisational level" and drive improvements.
These sections on disproportionality from the Inclusion Strategy at item 7a may be of interest.
Eg as reflected in arrests of children www.police-foundation.org.uk/publication/...
What I would add is that, at least in London, there are some quite clear differences in the profile of offending behaviour (eg crime types) by ethnicity, which we might expect to translate to some differences in outcomes.
These are the data for E&W across the three years. If you were to remove the verbal warnings, other action, and guardian interventions you'd end up with a 'positive outcome rate' (or whatever terminology is appropriate) of 28.2% Asian, 28.5% Black and 27.2% White.
I thought this was interesting. While stop and search 'criminality detected' rates vary between forces, at force level those rates vary little by subject ethnicity, especially in the large metropolitan forces. In other words, the bar for reasonable suspicion is applied quite consistently.
@mattashby.com this might be of interest
@stopwatchuk.bsky.social
Criminality detected = all outcomes apart from NFA.
The Home Office stop and search open data are in this file assets.publishing.service.gov.uk/media/6909d8... via www.gov.uk/government/s...
I thought this was interesting. While stop and search 'criminality detected' rates vary between forces, at force level those rates vary little by subject ethnicity, especially in the large metropolitan forces. In other words, the bar for reasonable suspicion is applied quite consistently.
It's strength is that it isn't party political
LPB happening now webcasts.london.gov.uk/Mayoral/Even...
So a model that can't explain all of the variance in disproportionality suggests that in East Sheen levels are 3-4x higher than expected.
And the model only explains 62% of the variance in the dispro data
Notably, the MOPAC modelling suggests 'expected' disproportionality in East Sheen would be x14
