Jose Pina-Sánchez 🇪🇺
banner
jpinasanchez.bsky.social
Jose Pina-Sánchez 🇪🇺
@jpinasanchez.bsky.social

Professor of Quantitative Criminology and co-director of the Social Research Methods centre at the University of Leeds. Interested in #Data #Bias #Measurement #CriminalJustice #Sentencing #Disparities.
jmpinasanchez.github.io/ .. more

Political science 34%
Sociology 29%

Came across this based on our latest paper: odysee.com/@crimconsort...
It is truly amazing how much better these LLMs are at explaining ones own ideas. I was recently given 90 minutes to present this paper, yet I bet had I played this 6 minutes clip that would have been way more effective.
Crimversations: "Estimating Discrimination in Sentencing: Distinguishing between Good and Bad Controls"
This is a podcast based on the article, "Estimating Discrimination in Sentencing: Distinguishing between Good and Bad Controls" (available at https://doi.org/10.21428/cb6ab371.3fba9264). We create "cr...
odysee.com
This (free) event may be of interest our members: “Towards new horizons of scholarly publishing” in Nijmegen on 5-6 February '26. Topics: future of academic publishing, Diamond Open Access, and challenges ahead. See: www.horizondiamond.nl
New Horizons Diamond
www.horizondiamond.nl
#AcademicSky #PsychSciSky

Not read yet, but this paper looks of interest to people in my orbit.

Questioning the practice of throwing every covariate under the sun into your models. Here, in a legal context, which is interesting.
#CriminalJustice and #Legal scholars working on #empirical topics, do submit your work here. I had a fantastic experience, thorough reviews, short waiting time, great communication with the editorial team, and it is all 💎OA, so my university doesn't have to send £3K to Springer Nature stakeholders.
'Estimating Discrimination in Sentencing: Distinguishing between Good and Bad Controls', with Melissa Hamilton & @pwgtennant.bsky.social, just out at the European Journal of Empirical Legal Studies.
publicera.kb.se/ejels/articl...

Shareholders
Fewer workers. Fewer international students. More people leaving.

A "step in the right direction" according to the PM.

Impossible to take PM/govt seriously on growth if they are deliberately reducing it (and making the fiscal position worse) *as a matter of policy*.

www.bbc.co.uk/news/article...
UK net migration falls sharply with drop in arrivals for work and study
Provisional figures for net migration to the UK show levels dropped to 204,000 in the year to June 2025.
www.bbc.co.uk

Reposted by Gordon Hodson

#CriminalJustice and #Legal scholars working on #empirical topics, do submit your work here. I had a fantastic experience, thorough reviews, short waiting time, great communication with the editorial team, and it is all 💎OA, so my university doesn't have to send £3K to Springer Nature stakeholders.
'Estimating Discrimination in Sentencing: Distinguishing between Good and Bad Controls', with Melissa Hamilton & @pwgtennant.bsky.social, just out at the European Journal of Empirical Legal Studies.
publicera.kb.se/ejels/articl...

In short, we need to move on from a positivist to a post-positivist perspective. Stop presenting findings as point estimates (which are always, irredemiably, going to be wrong) and instead offer a band of likely values.

However, that is not to say that judicial prejudice can never be detected. In certain cases, when that prejudice is present and the model uncertainty is narrow enough, we can. See example for the case of disparities against Black offenders.

We do not know what is the right set of controls, so the only honest answer is to find a way reflect that model uncertainty. This involves accepting that the true extent of judicial prejudice is unknowable. See model uncertainty in disparities in sentence length vs Hispanics in the US Federal Court.

We highlight a trade-off between confounding and post-treatment bias when it comes to deciding what to do with judicially-defined case characteristics like remorse. Criminologists tend to control for such case characteristics, while Econometricians dont. They are both wrong.

We look at sentencing, but what we discuss is relevant to other CJS processes and even other decision-making contexts where there is discretion in both the outcome and how the case is defined. We should be mindful that prejudice can manifest at either stage and choose our controls accordingly.

Reposted by Juanjo Medina

'Estimating Discrimination in Sentencing: Distinguishing between Good and Bad Controls', with Melissa Hamilton & @pwgtennant.bsky.social, just out at the European Journal of Empirical Legal Studies.
publicera.kb.se/ejels/articl...
80% of UK women have experienced public harassment, yet 75% report that no one stepped in.

Dr @annabarker.bsky.social's research has helped hundreds of park staff and volunteers across the UK to be trained in how to spot, safely challenge, and prevent harassment.

Read more👇
tinyurl.com/99px5urh

The same applies to time, absorbed by teaching and service, and access to networks too. It is not the same to carve your own path than to be led by the hand.

"Local universities are the top international exporting sector in cities including Exeter, Dundee, Leicester and Nottingham and are among the largest high skilled “knowledge” sector employers in many parts of the UK." www.centreforcities.org/press/univer...
To shit with exports and regional growth.

09.26 in the morning and I have already been asked to verify my OneDrive and Outlook account 5 times. No one does productivity like Microsoft.
Every time the BBC mentions the "black hole" in the UK public finances, they should say "due to Brexit".

And every time they ask someone about tax rises or spending cuts they should ask "was Brexit worth it for this?"

"Que factores influewn en las sentencias judiciaeles? Las sentencias penales como objeto de investigación criminológica", (www.youtube.com/live/94S9Al_...) desde la Universitat Oberta de Catalunya, con Laura Arantegui, Daniel Varona y Josep M. Tamarit.
Las sentencias penales como objeto de investigación criminológica
YouTube video by UOC - Universitat Oberta de Catalunya
www.youtube.com
HYBRID EVENT: 'Introduction to Quantitative Bias' with Rachel Hughes on Monday 1st December at 10am-12pm UK time.

For those near Leeds, bring your laptop and enjoy this as an in-person session!

Sign-up before 9am on Thursday 27th via: www.eventbrite.co.uk/e/introducti...
We're organising a webinar on 18 February to mark the end of my @adr-uk.bsky.social Fellowship on school absences and crime. With talks from top-notch colleagues including @iainbrennan.bsky.social, @foxnic.bsky.social, Mark Mon-Williams, and many others!
www.eventbrite.co.uk/e/from-schoo...
From School Absences and Exclusions to Crime: Evidence and Implications
This webinar marks the conclusion of an ADR UK Fellowship examining the longitudinal relationship between school absences and crime.
www.eventbrite.co.uk
U.K. universities have seen a drop in overseas students. Plus many have a budget shortfall, so the impact is on staff redundancies resulting in larger tutor groups, loss of non academic staff, loss of facilities, loss of spending in uni towns and cities, loss of innovation support locally.
i: Reeves to unveil £600m raid on foreign student
university fees #TomorrowsPapersToday

RCTs are the gold standard! They pyramid says so.

Losing the US is actually an opportunity.

Missing not at random.
The UK's anti-migrant mania is destroying the health system with a 26% rise in overseas-trained doctors quitting the UK in 2024

www.theguardian.com/society/2025...
Overseas-trained doctors leaving the UK in record numbers
Medical bodies warn that hostility towards migrants is behind a 26% rise in departures last year that imperils NHS
www.theguardian.com

You should write an impact case for the ref about this.

Perfect example for why open peer review should be the way forward.
Our paper, that we worked hard to produce before the second wave, estimated that 21,000 excess deaths were caused by the 1 week delay in lockdown.

The paper was held up & eventually rejected by the Lancet. By the time it came out it could only offer a post mortem.

journals.plos.org/plosone/arti...
The content is still obviously delusional to anyone reading it. It’s just organised crazy. But the real danger is the internal effect on the user. The AI provides coherence and validation, stabilising the narrative they’re caught in.