Carlos E Lourenco (Caê)

in version control, every change that is made from the start of the document through its eventual publication can be easily documented in a git repository (like on github or bitbucket). the commit history of a project repository describes the entire development of the manuscript, in perpetuity

in 2020 my lab quit Microsoft Word and began writing in R Markdown, an open-source document-publishing system that facilitates a clear, permanent link between statements in text (scientific prose) and the code producing analyses and visualizations supporting those statements (scientific analyses)

🔪"As research on topics such as stereotype threat and implicit bias were found to have problems, the scholars [...] didn’t change their policy views, but simply kept citing the old, now-discredited findings, or else just made the empirical claims without any citations"

Chess position recall by age and rating. The paper found age insignificant because it only looked at its linear effect, but recall drops after age 50 regardless of rating. #dataviz

“Which means that any time a traffic cop forgot to fill in the license plate number on a citation, the fine automatically got sent to Joseph Tartaro.”

DeepSeek DSpark for DeepSeek-V4 Flash and Pro - 60% to 85% faster generation? Models: huggingface.co/deepseek-ai/... Paper: github.com/deepseek-ai/... Repo: github.com/deepseek-ai/...

Everyone should read this, just to see how they implemented a perceptron (a very basic neural network) in Age of Empires II arxiv.org/abs/2605.31514 The artificial neurons pass activation energy in the form of goats.

Delighted that this paper with Kristina Butaeva @butaevak.bsky.social and Alexander Shapoval is now public.

Just found this intriguing new function by Wadea Abu Dahoud for making a log-scale axis squiggly, mostly to act as a visual cue/speed bump that the axis isn't linear. github.com/wade31985-ar... #rstats #dataviz

🔪"The social justice movement emphasizes social identity while simultaneously diminishing the mechanisms that can rein in the negative impacts of doing so."

I’ll happily refer you to the paper for the details :) projecteuclid.org/journals/sta...

gonna post 100 likely permutations of the World Cup knockout bracket on bsky and then delete all the ones that were wrong after the final and then claim I predicted everything before the tournament to demonstrate p-hacking for my classes next semester

🔪"For example, a widely reported 2023 study purported to undermine the broad scholarly consensus that almost all of the hunting in hunter-gatherer societies is done by men..." higheredstatementofprinciples.com/state-of-sch...

Claude Cowork doubled weekly usage limits until July 5th. Example of how I use Claude Cowork. I Built an Obsidian second brain for my Claude Cowork project 😎 ai.georgeliu.com/p/my-claude-...

Claude Cowork doubled weekly usage limits until July 5th. Example of how I use Claude Cowork. I Built an Obsidian second brain for my Claude Cowork project 😎

Very thoughtful interview with @margaritap.bsky.social on why social media bans in youth don’t achieve the aims people believe they will — and what to do instead. newhumanist.org.uk/articles/sho...

"A lack of care, a lack of regard for what is important, paired with a lack of regard for honesty and transparency in design, is what characterizes bullshit technologies and the design processes that produce them."

@dingdingpeng.the100.ci do you guys have merch? I need this on a t-shirt.

You shouldn't even need that for some packages, since they ensure that openMP works out of the gate if you install the MacOS binaries from CRAN. For example, fixest: github.com/lrberge/fixe...

See mac.r-project.org/openmp/ for full details

Installing stuff on a new M5 MacBook Pro and finally revisiting my old Makevars incantations and just learned that life is like a billion times easier nowadays for getting #rstats to work with OpenMP on macOS! Basically I just have this now and everything installs/compiles smoothly?

New blog post! Let's say an effect you are interested in varies between countries and you want to explain some of that cross-cultural variation. What could possibly go wrong? www.the100.ci/2026/05/20/f...

The FT article is gross. I was on board through the lack of affordable housing section, but then he goes on to say something like sub-Saharan African women have gotten unrealistic expectations for partners that local men aren’t prepared for and I was done.

Smartphones can do a lot, but apparently they cannot keep you from analysing your data in an incredibly wacky manner to feed a certain narrative about smartphones.

Simulating data russpoldrack.substack.com/p/simulating... The latest from my Better Code, Better Science series

Gemini CLI hacking example www.youtube.com/watch?v=74Fj...

CausalReasoningBenchmark is a new tool for automated causal inference that distinguishes between causal identification and estimation. With 173 queries, it demonstrates that LLM models struggle with nuanced research design, aiming to enhance rigor in causal analysis. https://arxiv.org/abs/2602.20571

Feed: "Statistical Modeling, Causal Inference, and Social Science" By: Bob Carpenter on Saturday, May 16, 2026

🧐 🔓 Dalla Bona, S., Navarrete, E., & Vicovaro, M. (2026). The foreign language effect in the illusion of causality: Evidence of absence. Bilingualism: Language and Cognition. Advance online publication. doi.org/10.1017/S136...

Synthetic control is extremely popular. But are the popular recommendations on how to use it actually valid? In this week's issue of Causal Python Weekly: 1/ #CausalSky #StatSky #EpiSky #EconSky

Meet the academics refusing to use generative AI Researchers say they have their reasons for avoiding AI tools — and they’re sick of arguing about it. www.nature.com/articles/d41... h/t @maartenp.bsky.social

This by @lucyavraamidou.bsky.social might be useful doi.org/10.1007/s111...

will you record the session and post it on the youtube channel (www.youtube.com/@theCIIG)?

A slide from this talk I'm giving on Monday. The conclusion will be a set of questions that one must be able to answer to ask associational research questions. The ball is in your court association people!

Jeremy right now

OpenAI rolls out a Chrome extension for Codex, allowing agents to run browser sessions independently; Additionally, tests Remote Control and Voice Mode!

Fabricated citations: an audit across 2·5 million biomedical papers The fabrication rate increased more than 12 times between 2023 and early 2026. Possible causes: Paper Mills, AI use etc www.thelancet.com/journals/lan...

What can we learn from automating an entire quantitative social science paper, from prompt to finished product? Thread about ongoing work with @natewilmers.bsky.social 1/12 Paper: osf.io/preprints/so...

Paper mill content growing faster than legitimate research, according to study by @reeserichardson.bsky.social & colleagues — which @elisabethbik.bsky.social mentioned in her talk at #WCRI2026: cen.acs.org/policy/publi...

The WiSo 💚 is ranked 4th🎉 among the 15 most productive marketing departments worldwide over the last six years, according to the Financial Times FT50 Marketing Journals. Germany-wide, the faculty achieves the best ranking. Congratulations to the Marketing Department!📢💐💚 Details: uni.koeln/BR6JK

In search of the ‘spirit of the SDGs’: New research assessment tool to assess relatedness of business school research published in FT50 journals to #SDGs #impact via Kathleen Rodenburg, @vinulidesilva @YU_President https://www.mdpi.com/1412054 #mdpisustainability via @Sus_MDPI

Great news that Psychological Science, an Association for Psychological Science journal, has been named in the Top 50 of influential journals by the Financial Times. 🎉 www.ft.com/ft50-journals @psychscience.bsky.social