Alex Sutherland
@criminologist.bsky.social
Criminology, public policy, research and experiments
I was struck by this stat re UK firearms officers the other day:
4 / 17,249 times where firearms officers were deployed where a firearm was shot (not Inc taser)
0.0002%
4 / 17,249 times where firearms officers were deployed where a firearm was shot (not Inc taser)
0.0002%
November 11, 2025 at 7:41 AM
I was struck by this stat re UK firearms officers the other day:
4 / 17,249 times where firearms officers were deployed where a firearm was shot (not Inc taser)
0.0002%
4 / 17,249 times where firearms officers were deployed where a firearm was shot (not Inc taser)
0.0002%
I think I could make a whole presentation using this.
November 9, 2025 at 9:28 AM
I think I could make a whole presentation using this.
Use of AI in education makes me think of this scene from Real Genius (1985!)
All it'll take is the lecturer using AI to ask the questions and the loop is closed.
Everyone loses.
All it'll take is the lecturer using AI to ask the questions and the loop is closed.
Everyone loses.
November 8, 2025 at 9:25 AM
Use of AI in education makes me think of this scene from Real Genius (1985!)
All it'll take is the lecturer using AI to ask the questions and the loop is closed.
Everyone loses.
All it'll take is the lecturer using AI to ask the questions and the loop is closed.
Everyone loses.
Study on art, funded by Art Fund, finds that there are health benefits from looking at, hmmmm, art, in art galleries.
Oh and that these benefits "are universal".
From n=50.
Got it.
Oh and that these benefits "are universal".
From n=50.
Got it.
November 4, 2025 at 7:51 AM
Study on art, funded by Art Fund, finds that there are health benefits from looking at, hmmmm, art, in art galleries.
Oh and that these benefits "are universal".
From n=50.
Got it.
Oh and that these benefits "are universal".
From n=50.
Got it.
Malmö in high street fashion (pretty random) but made me think of @mannegerell.bsky.social, so...hej!
November 1, 2025 at 12:13 PM
Malmö in high street fashion (pretty random) but made me think of @mannegerell.bsky.social, so...hej!
Can't believe I forgot about...
October 31, 2025 at 8:40 PM
Can't believe I forgot about...
Ok need to be stopped as this could go all night
October 31, 2025 at 2:56 PM
Ok need to be stopped as this could go all night
Again, what's the plan for AI increasing unemployment? Anyone?
AI will, what, be so efficient people don't have to work but get paid anyway?
(And when the bubble bursts this will be so, so bad)
AI will, what, be so efficient people don't have to work but get paid anyway?
(And when the bubble bursts this will be so, so bad)
October 28, 2025 at 10:35 PM
Again, what's the plan for AI increasing unemployment? Anyone?
AI will, what, be so efficient people don't have to work but get paid anyway?
(And when the bubble bursts this will be so, so bad)
AI will, what, be so efficient people don't have to work but get paid anyway?
(And when the bubble bursts this will be so, so bad)
Lunchtime discussion of pasteurisation Vs antibiotics in milk led to looking up "raw" milk [sic] supply in UK. Interesting health claims:
October 26, 2025 at 2:06 PM
Lunchtime discussion of pasteurisation Vs antibiotics in milk led to looking up "raw" milk [sic] supply in UK. Interesting health claims:
October 24, 2025 at 6:08 PM
Every time I see the MZ photo...
Also don't get what's in it for wearers but also don't care enough to find out.
Also don't get what's in it for wearers but also don't care enough to find out.
October 23, 2025 at 5:24 PM
Every time I see the MZ photo...
Also don't get what's in it for wearers but also don't care enough to find out.
Also don't get what's in it for wearers but also don't care enough to find out.
Probably
Not
Actually
Significant
-There is a lot of missing data.
-It varies differentially by treatment group and time point.
-unclear about timing of consent Vs knowledge of treatment group.
-no adjustment AFAICT for multiple outcomes/time points but haven't looked at SI docs yet.
Not
Actually
Significant
-There is a lot of missing data.
-It varies differentially by treatment group and time point.
-unclear about timing of consent Vs knowledge of treatment group.
-no adjustment AFAICT for multiple outcomes/time points but haven't looked at SI docs yet.
October 22, 2025 at 5:02 PM
Probably
Not
Actually
Significant
-There is a lot of missing data.
-It varies differentially by treatment group and time point.
-unclear about timing of consent Vs knowledge of treatment group.
-no adjustment AFAICT for multiple outcomes/time points but haven't looked at SI docs yet.
Not
Actually
Significant
-There is a lot of missing data.
-It varies differentially by treatment group and time point.
-unclear about timing of consent Vs knowledge of treatment group.
-no adjustment AFAICT for multiple outcomes/time points but haven't looked at SI docs yet.
Saw this on linkedin re mobile phone theft and resale.
Massive effort, vast cost.
wouldn't be necessary if Apple (and it's mainly iphones) prevented access to its cloud worldwide once a phone was reported stolen.
At present it's only in UK for UK registered phones.
Massive effort, vast cost.
wouldn't be necessary if Apple (and it's mainly iphones) prevented access to its cloud worldwide once a phone was reported stolen.
At present it's only in UK for UK registered phones.
October 22, 2025 at 8:19 AM
Saw this on linkedin re mobile phone theft and resale.
Massive effort, vast cost.
wouldn't be necessary if Apple (and it's mainly iphones) prevented access to its cloud worldwide once a phone was reported stolen.
At present it's only in UK for UK registered phones.
Massive effort, vast cost.
wouldn't be necessary if Apple (and it's mainly iphones) prevented access to its cloud worldwide once a phone was reported stolen.
At present it's only in UK for UK registered phones.
And this is a great example of misattribution
Implied logic:
knife offence=more likely to be a murderer
But there are 50k knife offences a year so knife homicides were a 0.0049% of all knife offences
Implied logic:
knife offence=more likely to be a murderer
But there are 50k knife offences a year so knife homicides were a 0.0049% of all knife offences
October 19, 2025 at 3:41 PM
And this is a great example of misattribution
Implied logic:
knife offence=more likely to be a murderer
But there are 50k knife offences a year so knife homicides were a 0.0049% of all knife offences
Implied logic:
knife offence=more likely to be a murderer
But there are 50k knife offences a year so knife homicides were a 0.0049% of all knife offences
It's very common for fear of crime to increase as people age.
Also the chances of actual victimisation are much lower as people get older.
(I don't know who this person is but it prompted these thoughts.)
Also the chances of actual victimisation are much lower as people get older.
(I don't know who this person is but it prompted these thoughts.)
October 19, 2025 at 3:28 PM
It's very common for fear of crime to increase as people age.
Also the chances of actual victimisation are much lower as people get older.
(I don't know who this person is but it prompted these thoughts.)
Also the chances of actual victimisation are much lower as people get older.
(I don't know who this person is but it prompted these thoughts.)
"I just see what's going on...ONLINE in a weird personal bubble that has been created for me by algorithms".
- fixed it.
And if he needs a watch I can get him a decent Casio for like £20
- fixed it.
And if he needs a watch I can get him a decent Casio for like £20
October 16, 2025 at 8:14 AM
"I just see what's going on...ONLINE in a weird personal bubble that has been created for me by algorithms".
- fixed it.
And if he needs a watch I can get him a decent Casio for like £20
- fixed it.
And if he needs a watch I can get him a decent Casio for like £20
Decorative gourd season
October 15, 2025 at 9:06 PM
Decorative gourd season
I'm still wondering whether I'm supposed to be impressed or skeptical about the sample size. The provenance etc is what it is.
And as to the earlier point about 19 studies - they extracted data from 19 different studies. It's not like a combination of data from those studies.
so: N=87
And as to the earlier point about 19 studies - they extracted data from 19 different studies. It's not like a combination of data from those studies.
so: N=87
October 13, 2025 at 6:37 AM
I'm still wondering whether I'm supposed to be impressed or skeptical about the sample size. The provenance etc is what it is.
And as to the earlier point about 19 studies - they extracted data from 19 different studies. It's not like a combination of data from those studies.
so: N=87
And as to the earlier point about 19 studies - they extracted data from 19 different studies. It's not like a combination of data from those studies.
so: N=87
Bit of early morning Oxford.
October 9, 2025 at 7:05 AM
Bit of early morning Oxford.
Damn.
(I used to ❤️ MASK as a kid. I'm sure it's aged well...)
(I used to ❤️ MASK as a kid. I'm sure it's aged well...)
October 1, 2025 at 10:12 AM
Damn.
(I used to ❤️ MASK as a kid. I'm sure it's aged well...)
(I used to ❤️ MASK as a kid. I'm sure it's aged well...)
Weird to see people denying there was a 50% HE participation target set by Labour in early 2000s.
In this speech: news.bbc.co.uk/1/hi/uk_poli...
Tony Blair said:
In this speech: news.bbc.co.uk/1/hi/uk_poli...
Tony Blair said:
September 30, 2025 at 2:43 PM
Weird to see people denying there was a 50% HE participation target set by Labour in early 2000s.
In this speech: news.bbc.co.uk/1/hi/uk_poli...
Tony Blair said:
In this speech: news.bbc.co.uk/1/hi/uk_poli...
Tony Blair said:
news.bbc.co.uk/1/hi/uk_poli...
TB definitely said "this is a target" - so it's a target whether or not anything else happens.
TB definitely said "this is a target" - so it's a target whether or not anything else happens.
September 30, 2025 at 2:37 PM
news.bbc.co.uk/1/hi/uk_poli...
TB definitely said "this is a target" - so it's a target whether or not anything else happens.
TB definitely said "this is a target" - so it's a target whether or not anything else happens.
September 30, 2025 at 2:34 PM
Finished reading David Hand's Dark Data.
A very good introduction to missing data and the problems it brings in policy, practice and life in general.
Worth having on the bookshelf.
(One thing I can't believe I didn't know was randomised response for collecting sensitive self-report data. Neat!)
A very good introduction to missing data and the problems it brings in policy, practice and life in general.
Worth having on the bookshelf.
(One thing I can't believe I didn't know was randomised response for collecting sensitive self-report data. Neat!)
September 28, 2025 at 8:32 AM
Finished reading David Hand's Dark Data.
A very good introduction to missing data and the problems it brings in policy, practice and life in general.
Worth having on the bookshelf.
(One thing I can't believe I didn't know was randomised response for collecting sensitive self-report data. Neat!)
A very good introduction to missing data and the problems it brings in policy, practice and life in general.
Worth having on the bookshelf.
(One thing I can't believe I didn't know was randomised response for collecting sensitive self-report data. Neat!)