Scenes from the Dubai floods

Last month I gave a talk: can chatGPT pass my statistics module? exploring how we might use GPT's data analyst in teaching stats. I've now written that up in a blog post: rekadata.net/blog/2024-04-1… I am no expert, only a curious novice, so feedback and alternative opinions welcome!

I wrote this short blurb about Michael J Fox (Michael J. Fox) for the annual #Time100 . This man means a lot to me. And I know he means a lot to you… ⬇️

I know Michael J. Fox. Like, I know him pretty damn well. He’s funny. He’s warm. He’s handsome and intensely smart. He also

No. Smacking children should be banned because it’s violence, and violence is (almost always) immoral. This isn’t a scientific or medical question - it’s a moral one!

4+ years ago, I wrote about Buffalo's playbook to reduce street violence. Encouraging to see it's stuck with this evidence-informed blend of policing and community-led strategies. This story emphasizes the latter; I focused on the former.

thetrace.org/2024/04/buffal…

We propose new confidence intervals for Fuzzy RDDs. They take local linear regression bias explicitly into account, are short in canonical settings, but also valid in challenging ones with discrete running variables, donut holes, or weak identification econometricsociety.org/publications/e…

All that said, it seems intuitive that higher workloads will result in slower investigations and a generally poorer service to victims and the criminal justice system, and the main research finding seems to provide strong empirical evidence for that.

New research from Tom Kirchmaier and Ekaterina Oparina looking at high risk DV cases in GMP between 2014 and 2019: 'Higher workload for police officers correlates with victims withdrawing statements' blogs.lse.ac.uk/politicsandpol… 🧵

Zoom Link for my PhD viva, where I will defend my doctoral thesis titled “Support and treatment for men convicted of sexual offending: Readiness, change, and previous help-seeking”, on the 31st of May at 1.15 PM.

‘Brig. Gen. Reem Aminoach, former financial advisor to the IDF chief of staff, tried to estimate. 'The defense tonight was on the order of 4-5 billion shekels,' he said’ (about $1bn) ynetnews.com/article/h16o8q…

Sweden: A 39-year-old father was fatally shot while accompanying his 12-year-old son to the swimming pool in the socially disadvantaged area of Skärholmen. The incident occurred after the father reportedly attempted to disarm a group of youths. expressen.se/nyheter/sverig…

Läs Amirs tråd.

Just nu vill alla visa handlingskraft och göra något åt gäng och gängvåld.

Men vi behöver göra saker som fungerar snarare än saker som visar handlingskraft.
Och då behöver vi också ta reda på vad som faktiskt fungerar.

CCJ folks - what journals do we have that publish methods papers?

Säkerhetszoner - Bu eller bä?

Jag går igenom en del relevant forskning, och killgissar slutligen att svaret varken blir bu eller bä, utan snarare meh.

Facit får vi förhoppningsvis om något år när det utvärderats.

mgerell.blogspot.com/2024/04/sakerh…

People often think the most important part of science is the statistical analysis.

But they’re wrong.

The most important part is understanding the data. Who collected it, when, where, why, how, and from whom? Who entered and cleaned it, when, where, why, how, and for whom?

Early View in Criminology & Public Policy

“Attempting to reduce traffic stop racial disparities: An experimental evaluation of an internal dashboard intervention”

Open Access for all to read. A 🧵below: doi.org/10.1111/1745-9…

1/10

The lesson here is this:

When a politician says they’re going to create a “specific offence” to cover behaviour that’s already criminal, you can assume they’re doing it so they can say they’ve ‘done something’, usually because actually fixing the problem is too hard/expensive.