Ryan Larson

New! How homicide changed in Minnesota’s Twin Cities after the police murder of George Floyd. Read @ryanplarson’s thread (and data viz) and/or our team’s new article in J of Quant Crim.

Overall, we document an enduring, compositionally distinctive, and spatially concentrated spike in homicide following Floyd’s murder, partially mediated by de-policing. This highlight how institutional ruptures can interact with chronic inequality to reshape lethal violence.

Fourth, we observe a decrease in MPD stops after the murder, and reduced proactive policing explains a meaningful portion (~25%) of the post-murder increase in homicide, suggesting that both de-policing and broader social and institutional mechanisms were also at work. 5/n

Fresh online in JQC (@SpringerCrim) w/ @jillkpeterson, @chrisuggen, and others! We use an ITS design to examine the changes in the Twin Cities in relation to the pandemic and the police murder of George Floyd. Finding summaries below! Link here: link.springer.com/article/10.100… 1/n

"Breaking Bad: How Health Shocks Prompt Crime" aeaweb.org/articles?id=10…

"Can you explain how p-value is calculated for your synthetic control model? Me:

Juvenile Arrest as a Life Course Trap: Being arrested as a juvenile is associated with a 20-30 percentage-point drop in the likelihood of graduating from a 4-year college. -Garrett Baker, David Kirk, & Rob Sampson, new in SocIology of Education journals.sagepub.com/doi/full/10.11…

Comparing group means with t-tests is a common task in applied statistics, but the underlying logic is often hidden behind one-line function calls. infer enables tidy t-tests in R by expressing hypothesis testing as a readable, step-by-step workflow that fits naturally into tidyverse pipelines. Useful when you want to: ✔️ Express t-tests as a clear sequence of inferential steps ✔️ Visualize null distributions and observed statistics ✔️ Emphasize reasoning and uncertainty rather than formulas ✔️ Keep hypothesis testing readable and consistent with tidyverse workflows The visual below illustrates typical outputs from one-sample and two-sample t-tests using infer, taken from the official infer documentation. More details are available here: infer.netlify.app I regularly share practical tools and examples on statistics, data science, R, and Python in my newsletter. Click this link for detailed information: eepurl.com/gH6myT #datascienceenthusiast #rstudioglobal #Rpackage #tidyverse #RStats

We're mourning yet again in Minnesota — and struggling to square the gut-churning video of Alex Pretti's final moments with official accounts. I teach about techniques of neutralization, so I made a simple slide on the DHS statement. Share if it's useful. nytimes.com/2026/01/24/us/…

The need is no doubt greater than the increase in call volume, since the federal crackdown likely increased legal cynicism and system avoidance, both of which reduce the likelihood of calling authorities for help.

There are many signs of the fear and distress people are feeling in my home town of Minneapolis. One signal of elevated strain is the rise in 911 calls for help — up about 9% daily in the new year relative to previous daily averages across 2020-2024.

Now out in Nature Human Behaviour! 🚀🚀 Over the past decades, research on collective human behaviour has relied heavily on networks. This is intuitive: people interact with other people. However, we argue that this dominant framework misses a crucial ingredient. Traditional networks represent agents as nodes and pairwise relations as edges. As a result, they fundamentally assume that social interactions can be decomposed into pairs. Yet many social processes are irreducibly group-based. A simple example: a group of three coauthors writing a paper cannot be reduced to three independent pairs of coauthors. The group itself matters. In this article, we review a wide range of empirical and theoretical cases where group interactions cannot be decomposed into pairwise ones, and show that higher-order interactions shape collective behaviour above and beyond dyadic ties. We advocate studying collective behaviour on hypergraphs, where interactions can involve multiple agents simultaneously. We review how hypergraphs provide new insights across domains, including affiliation and collaboration networks, high-frequency contact settings (families, friends), and key social processes such as social contagion, cooperation, truth-telling, and moral behaviour. Finally, we outline promising directions for future research: addressing computational challenges of higher-order models; studying bias and inequality in group dynamics; combining hypergraphs and large language models to investigate the coevolution of language and behaviour; and using higher-order networks to simulate the impact of policies before implementation; and others. We are very excited about this work and hope it will inspire further research in a rapidly growing and fundamental area with broad real-world implications. Link to the paper in the first reply This work was brilliantly led by Federico Battiston (@fede7j), with an outstanding team of co-authors: Fariba Karimi (@fariba_k), Sune Lehmann, Andrea Bamberg Migliano, Onkar Sadekar (@OnkarSadekar), Angel Sanchez, & Matjaz Perc (@matjazperc)

Indiana Jones is a fantasy about being a professor not having to submit itemized travel expenses after every adventure.

1/ Can a Google search hurt your chances of getting a job more than a background check? A new study with Sarah Lageson on criminal records, race, and willingness to hire has surprising answers. 🧵

📊I've uploaded the 3rd stats lab post: "What is the Logic Behind a Two-Way Fixed Effects Analysis?" It includes the meaning of control unit or/and time fixed effect with graphs and R codes. #twfe" target="_blank" rel="nofollow noopener">idlhy0218.github.io/page%20buildin… #rstats #econtwitter #soctwitter #dataviz

New! Survey experiment & interview study shows how narratives can reduce stigma of criminal via moral justification, social affinity signaling & information salience. -@DJHardingSoc @mariassmithphd @_daeunjung S Luna-Lopez & @PandaGlazer in @CulturalSociol link.springer.com/article/10.105…

High-frequency location data show that race affects citations and fines for speeding | Science science.org/doi/full/10.11…