Felix Bethke

New article: “The Language of Responsibility in the United Nations Security Council, 1946–2020” with @felixhaass & @holger_niemann in @ISQ_Jrnl Article: doi.org/10.1093/isq/sq… Preprint: felixhaass.de/www/manuscript… Data/Code: t.ly/ChKOG

Laying off 15% of my Claude code agents due to AI

@exQUIZitely Bundesliga Manager Professional ftw! Baunatal was allways the main rival.

This one goes out to all German-speaking retro gamers. I'll take an educated guess and say most of you will have played at least one, if not all, of these absolute classics and benchmark-setting football manager games back in the day - right? My all-time favourite is Bundesliga Manager Hattrick. Honourable mention goes to Starbyte's Super Soccer, which was actually created by a fan as a solo project (originally called Kicker, if I recall correctly) and then licensed by Starbyte for being totally awesome. What's your #1 football manager game?

🚨Replication alert🚨 I'm pleased to announce that my replication of Moretti (2021) is now accepted as a comment at AER. I find ten issues in the paper. My comment focuses on two major problems; in the appendix, I document eight (relatively) minor problems. 1/

Matjes roll - salt cured herring with lettuce and onion at FC Hansa Rostock (@hansarostock, Germany) 💶 €3 (£2.50)

Currywurst & Pommes at SV Alemannia Waldalgesheim, Germany 💶 €5.50 (£4.70)

When you collect data online, are the results from humans or AI? In a project led by Booth PhD student Grace Zhang, we estimate the prevalence of AI agents on commonly used survey platforms: osf.io/preprints/psya… 🧵

I just gave a closed-book, pen-and-paper midterm exam in my 300-level course at UBC with 100 students. All exams were graded by an experienced graduate-level TA according to a rubric. *** The average was 64/100.*** My class averages at UBC are usually 80-85. Context: • This was the first midterm, covering ONLY 4 weeks of material. • Students had a list of possible questions in advance: no surprise questions. • Questions included (a) 3 concept definitions, (b) 3 paragraph-long questions, and (c) a 1.5-page essay. • I have taught this class multiple times. Nothing in my teaching style changed this semester. • We read entire paragraphs of text in class, so students don't have to do something on their own that wasn't covered during the lecture. • Students take a 10-question multiple-choice quiz at the end of every class (30% of the final grade). • Attendance is 95-99% every class. Attention during lectures and participation in pair-work activities are very high → anticipating the end-of-class quiz. *** But unfortunately, I suspect many students are not reading the material on the syllabus. They are asking LLMs to summarize it instead.*** After the midterm, students reported: • They thought they knew concept definitions but couldn't produce them on paper. • They thought they understood the arguments but struggled to connect them or identify points of agreement and disagreement. My view: It might be “cool” or “innovative” to teach students to summarize readings with ChatGPT or write essays with Claude. But we may be doing them a disservice: reducing their ability to retain material, think creatively, and reason from what they know. If you only read what AI has summarized for you, you don’t truly "know" the material. Moving forward: We have a second midterm coming up. I don't know how to convey to students that the best way to do better on the exam is to rely on and improve their own reading skills.

Contrary to "widespread perception", Wikipedia and Grokipedia have "broadly comparable tendencies ... in their treatment of politically controversial topics", but there is "a modest but consistent right-leaning bias in Grokipedia relative to Wikipedia" arxiv.org/pdf/2601.15484

Social scientists working with materials requiring digitization can only study what machines can read. In practice, that means printed Latin-script documents from well-funded archives. In a new working paper, I show that Vision Language Models used zero-shot outperform every existing OCR system across every script evaluated, and I propose a pipeline for deploying them on new collections. I apply it to six archival collections spanning 1.8 million pages across six countries for under $1,900.

We just released #OpenRefine 3.10.1, with the Mac build correctly notarized. If you don't use a Mac, you can ignore this release. github.com/OpenRefine/Ope…

omg this title, this paper

I come from a country with a secret nuclear armament program. This project is deeply personal. The dataset took 10 years, digging through undigitized archives & interviewing UN staff who witnessed Africa's disarmament in the 80s. 40 years of DDR. 407 program-years. 57 variables. Trying to explain why some armed groups *willingly* enter DDR but rearm during the program. doi.org/10.1177/002200…

OpenRefine Wikimedia CommonsExtension v0.1.5 released. This update ensures all calls to Wikimedia services include a proper User-Agent header, enabling correct authentication and login to Wikimedia Commons. github.com/OpenRefine/Com… #OpenRefine #Wikimedia #OpenSource

@chris_dworschak @CCEW_Munich Congratulation to @CCEW_Munich. Chris is the best!

Our Coup D’Etat Project has a new data dashboard! Go check it out here: buff.ly/3ClMTgC

🚀 OpenRefine 3.10.0 is officially released! 🎥 We’ve also prepared a video walkthrough showcasing the main updates and how to use the new features in practice. 👉 github.com/OpenRefine/Ope… Huge thanks to everyone in the community who contributed code, testing, and feedback

I'm starting to think the problem is Thomas...

🚨 Interested in survey experiments? I'll be giving a virtual talk on (false) null results in experiments--and how to protect against them--at SWERP on 3/13 (3/12 8pm EST). Hope to see you there, and thanks to @sysilviakim & @_Brandon_Ives_ for inviting me!

This was one of the greatest football games ever 👌

Citation managers such as Zotero etc are more annoying than they are useful and instead of using them, it's better to just manually compile bibliographies