nicolaiberk.bsky.social - Profile | ThreadSky | a Reddit-style client for Bluesky

nicolaiberk.bsky.social

Postdoc at ETH Public Policy Group & Immigration Policy Lab. Political Communication, Public Opinion, NLP.

40 posts 738 followers 202 following

Posts 22 Comments 27

Curious about mastering text analysis using computational methods? 🧑‍💻📊 This summer, I’m teaching an introductory course at the @methodsnet.bsky.social Summer School at @weareceu.bsky.social in Vienna. Would be nice to see you there!

submitted 1 day ago • 0 comments

Ich muss es so hart sagen: ich dachte, ich lese eine kleine Anfrage der AfD, aber sie stammt von der Union. 551 (!) Fragen zu Organisationen, Hintergrund seien „Proteste gegen die CDU“, darunter die „Omas gegen Rechts“ und andere. Bin ehrlich entsetzt. dserver.bundestag.de/btd/20/150/2...

submitted 2 days ago • 402 comments

‼️ O, Canada is now in CHES! The data include party position estimates for federal parties in Canada + parties in Ontario and Quebec. It's also super easy to combine with other @chesdata.bsky.social data to compare positions with positions of parties in other countries!

submitted 7 days ago • 0 comments

www.readthedetox.com/p/trump-can-...

submitted 11 days ago • 0 comments

Wir starten das Justiz-Projekt. Wohin man schaut, stehen Attacken auf die Justiz im Mittelpunkt der autoritär-populistischen Strategie. Im Justiz-Projekt fragen wir uns jetzt: Wie verwundbar ist die deutsche Justiz? Alle Infos in unserem aktuellen Editorial: verfassungsblog.de/wir-starten-....

submitted 13 days ago • 5 comments

Dystopian af

submitted 13 days ago • 0 comments

🚨 #Data alert 📣 We just released #ParlLawSpeech – full texts of more than 40k bills, 28k laws, and 3 mio. parliamentary speeches from 7 countries (AT, CZ, DE, DK, ES, HR, HU) and the EU! If you study democracy with #TextAsData / #NLP methods, this is for you! A short 🧵 (1/3) #PoliSkyData #polisky

submitted 14 days ago • 11 comments

Heute in einer Woche, also wenige Tage vor der Bundestagswahl, darf ich zusammen mit @lionw.bsky.social & Tanja Blom in der Bezirkszentralbibliothek in Tempelhof über die vorgezogene Wahl diskutieren. Moderiert wird die Diskussion von @heikekleffner.bsky.social.

submitted 15 days ago • 1 comment

submitted 15 days ago • 1 comment

The much sympathy I have for this view, this is fatalistic. There is urgency to resist the turn into fully-fledged authoritarianism now. Resistance in authoritarianism is much harder cause things you take for granted right now in your repertoire like freedom of speech or assembly cease to mean much.

submitted 16 days ago • 1 comment

Steve Levitsky, co-author of How Democracies Die "the U.S. is sliding toward a more 21st-century model of autocracy: competitive authoritarianism" www.theatlantic.com/ideas/archiv... "A failure to resist... could pave the way for authoritarian entrenchment" www.foreignaffairs.com/united-state...

submitted 17 days ago • 11 comments

Die Linke bei den 18-29-Jährigen bei forsa bei 19%. (Auch wenn die Fallzahlen hier sicherlich nicht sehr hoch sind.)

submitted 17 days ago • 10 comments

This afterthought - to a 🧵 about a great publication, kudos @nicolaiberk.bsky.social - makes a crucial point. Making PhD degrees contingent on published papers hijacks candidates. Papers in dissertations should be ready for submission. Accepted papers are wonderful but should not be required.

submitted 18 days ago • 6 comments

Speaking of null results, have a look at this outstanding study by @nicolaiberk.bsky.social, just out with us at Political Communication, where he shows with exemplary rigor how BILD's sudden shift towards framing immigration as related to crime had zero effect on readers' attitudes on immigration.

submitted 17 days ago • 1 comment

I was just told that I have to remove “climate” from the title of an ongoing grant if I want to keep it. And publications from that grant cannot include “climate” and other forbidden words. I can’t believe I’m writing this from the United States of America. #AcademicSky

submitted 17 days ago • 402 comments

Fascinating article making use of an unexpected change in editor at Germany’s largest tabloid to test whether a negative shift In how immigrants are portrayed by a newspaper produces a negative soft on readers’ attitudes. Conclusion? It did not.

submitted 18 days ago • 3 comments

New research by @nicolaiberk.bsky.social using GLES data 👇

submitted 19 days ago • 1 comment

Now available Open Access: 📰 Does (immigration) framing influence public opinion? 🧵 www.tandfonline.com/doi/full/10....

submitted 19 days ago • 4 comments

Yes, just say the causal statement and acknowledge that there are strong assumptions required to believe it

submitted 29 days ago • 0 comments

To govern LLMs, we need to force transparency about the training process - and potentially even whitelist specific data for training.

submitted 31 days ago • 1 comment

The morale of DeepSeek r1's story is clear: It is very obviously possible to squeeze a model into a very particular response setting for very specific queries EVEN WITHOUT A SYSTEM PROMPT. This means: We cannot trust any model whose dataset and training is not fully transparent.

submitted 31 days ago • 1 comment

📣 Excited to welcome the fabulous @dianebolet.bsky.social at the Politics Lecture Series @humboldtuni.bsky.social and @dynamics.bsky.social tomorrow! Come & join us!

submitted 32 days ago • 0 comments