joseph francis You're going to make me dust off this research note I care about methods and I spend a lot of my free time trying to write open source code to make more robust methods easier to use. Randomization-based RD is just not an example of a better method IMO

BREAKING: wars devastate economies.

"The Economic Consequences of War" nber.org/papers/w34389

Bayesian probability, like frequentist probability, is a model-based activity that is mathematically anchored by physical randomization at one end and calibration to a reference set at the other statmodeling.stat.columbia.edu/2025/10/20/bay…

New paper: LLMs are increasingly used to label data in political science. But how reliable are these annotations, and what are the consequences for scientific findings? What are best practices? Some new findings from a large empirical evaluation. Paper: eddieyang.net/research/llm_a…

Finding 1: LLM annotations show pretty low intercoder reliability with the original annotations (coded by humans or supervised models). Perhaps surprisingly, reliability among the different LLMs themselves is only moderate (larger models better).

Finding 2: This disagreement has significant downstream consequences. Re-running the original analyses with LLM annotations produced highly variable coefficient estimates, often altering the conclusions of the original studies.

We also developed a new R package, localLLM (CRAN.R-project.org/package=localL…), that enables reproducible annotation using LLM directly in R. More functionalities to follow!

github.com/deepseek-ai/De…

New data drop! 📊 A new extension of the Global Tariff Database is now available, covering the U.S. trade war (2018–2025) 🇺🇸🌏 It includes bilateral tariffs — those imposed by the U.S. and those faced by U.S. exporters — tracking all changes from Jan 2018 to mid-Aug 2025.

I am on the faculty job market this year! I am seeking tenure-track faculty positions to drive my research agenda on rigorous AI evaluation for science and policy. I am applying broadly across disciplines, and would be grateful to hear of relevant positions. Materials: 🧵

We’re releasing the full FinePdfs source code — plus new datasets and models! 🚀 📚 Datasets: • OCR-Annotations — 1.6k PDFs labeled for OCR need • Gemma-LID-Annotation — 20k samples per language (annotated with Gemma3-27B) 🤖 Models: • XGB-OCR — OCR classifier for PDFs

We are hosting virtual Replication Games on Friday November 13th, 2025 with UK Reproducibility Network @ukrepro.bsky.social. This is our 2nd year collaborating with UKRN. Psych, public health, pol sci and econ studies will be reproduced! Register here: surveymonkey.ca/r/I4R_Replicat…

LoL

Causal Inference in High-Dimensional Generalized Linear Models with Binary Outcomes. arxiv.org/abs/2510.16669