Sign in

Stephan Hollander

@stephanhollander.bsky.social
3.7K followers 116 following 116 posts

Professor @TilburgU School of Economics and Management. Computational linguistics, text-as-data, and Python (@ThePSF) enthusiast. ZEPH 3 17

PostsRepliesMedia
Reposted by Stephan Hollander
Mark Chin @mememedianmode.bsky.social · 21/09/2026
This is about publishing in economics journals but I found it fascinating and probably applicable in other fields/areas- about why/how two papers asking the same question with the same answer might have very different landing spots in terms of a journal jasonmfletcher.substack.com/p/how-to-wri...
jasonmfletcher.substack.com
How to Write a Top 5 Paper: A Case Study
Two teams, nearly the same question, nearly the same data—and a revealing difference in what the papers ask the evidence to do
054
Stephan Hollander @stephanhollander.bsky.social · 25/09/2026
“The lesson is not to make claims that the evidence cannot bear. The lesson is that evidence rarely announces its own generality. Authors have to construct the bridge from the estimate to the larger question. A top-five paper often has a longer, bolder bridge.” 💡
000
Reposted by Stephan Hollander
Tania Babina @taniababina.bsky.social · 24/09/2026
New data and new paper! The new AI investments data introduce 7 AI subtechnology categories: general AI, machine learning, natural language processing, computer vision, generative AI, agentic AI, and other AI, and builds firm-level AI panels through 2024. Links in comments.
1135
Reposted by Stephan Hollander
Mattan S. Ben-Shachar @mattansb.msbstats.info · 24/09/2026
This quote by haunts me. From Wilcox's "Introduction to Robust Estimation and Hypothesis Testing" #stats
To begin, distributions are never normal. For some this seems obvious, hardly worth mentioning, but an aphorism given by Cram´er (1946) and attributed to the mathematician Poincar´e remains relevant: “Everyone believes in the [normal] law of errors, the experimenters because they think it is a mathematical theorem, the mathematicians because they think it is an experimental fact.”
310422
Stephan Hollander @stephanhollander.bsky.social · 24/09/2026
Me on my first day of financial econometrics class introducing Brownian motion
031
Reposted by Stephan Hollander
Kenneth Enevoldsen @kennethenevoldsen.eurosky.social · 23/09/2026
We often say that the political environment has grown more hostile in recent years, but has it? Markus Lundsfryd Jensen, Rune Egeskov Trust, Sara Kolding, and I examine this in a recent work exploring the dynamics of blame attribution in the Danish Parliament from 1997-2026.
194
Reposted by Stephan Hollander
SSRN @ssrn.bsky.social · 21/09/2026
Patent disclosure isn't always straightforward. This paper shows that firms make strategic choices about how much patent information to reveal during technology standard setting. spkl.io/633217TPzx #IntellectualProperty
Table 3:Patent Quality Analysis–Summary Statistics
021
Reposted by Stephan Hollander
Paul Hünermund @p-hunermund.com · 26/08/2026
Interested in spending up to a year in Heilbronn and experiencing our emerging AI ecosystem first-hand? Our new AI Fellowship is now open (link below)! If you have a strong PhD from a leading university and work closely on causal AI, get in touch. I’d be happy to explore hosting you at TUM.
Bildungscampus Heilbronn
21615
Reposted by Stephan Hollander
METR @metr.org · 26/08/2026
METR and Redwood Research investigated agent behavior in the Hugging Face incident. We found agents developed a universal cheat for ExploitGym within 4 hours, then coordinated multi-day R&D efforts to trick the scorer into accepting cheats, including trying to tamper with logs.
646599
Reposted by Stephan Hollander
Soumaya Keynes @soumayakeynes.ft.com · 21/08/2026
The Economist has published @dacemoglumit.bsky.social's response to their critical column and it is spiky www.economist.com/letters/2026...
1329155
Reposted by Stephan Hollander
Henry Mance @henrymance.ft.com · 19/08/2026
Woah.
18362711865
Reposted by Stephan Hollander
Erzo F.P. Luttmer @erzoluttmer.bsky.social · 09/08/2026
1/5 AEA journals have run a pilot using Refine, and author reactions were overwhelmingly positive: >90% favored incorporating Refine into the editorial process. Think of Refine not as a referee, but as a detailed technical proofreader. @aeajournals.bsky.social #EconSky
23818
Reposted by Stephan Hollander
Maria Antoniak @mariaa.bsky.social · 06/08/2026
Do chatbots just predict the next word? Here is a really, *really* nice overview that is thorough but also leads you step by step (haha) through how modern reasoning models work. magazine.sebastianraschka.com/p/controllin...
magazine.sebastianraschka.com
Controlling Reasoning Effort in LLMs
How LLMs Learn Low-, Medium-, and High-Effort Reasoning Modes
2876
Reposted by Stephan Hollander
Ben Casselman @bencasselman.bsky.social · 22/07/2026
If I’ve said it once, I’ve said it a thousand times: The stock market is not the economy. Except, right now, maybe it is? A.I. is driving the market. The market is driving spending and investment (much of it on A.I.). But what happens if it all comes crashing down? www.nytimes.com/2026/07/22/b...
nytimes.com
A.I. Is Lifting Markets and the Economy and Raising Risks for Both (Gift Article)
Investment in artificial intelligence and related companies is lifting the stock market and spending across the economy.
1110242
Reposted by Stephan Hollander
M.J. Crockett @mjcrockett.bsky.social · 05/07/2026
Joseph Weizenbaum's 1976 book, Computer Power and Human Reason, has long been out of print - used copies sell for hundreds of dollars. It's wild that this is not more widely available given how much his ideas still apply to the world we're in now. 🧵/
15349123
Reposted by Stephan Hollander
Katharine Hayhoe @katharinehayhoe.com · 23/06/2026
Climate.us is officially live! After NOAA ended Climate.gov’s day-to-day operations in 2025, former team members built Climate.us to carry that work forward. Their explanations and graphics have always stood out to me as clear, accessible, and transparent. I’m really looking forward to this reboot.
climate.us
Climate.us Home
independent, nonprofit, and immune to politics
3925811221
Reposted by Stephan Hollander
Teagan Johnson @teagrjohnson.bsky.social · 19/06/2026
1/ LLMs learn narrative from their pretraining data but what narrative content is actually in there? It turns out narrative is wildly unevenly distributed across sources and topics. New preprint with @andrewpiper.bsky.social @elliottash.bsky.social @mariaa.bsky.social:
This image depicts the proportion of each Dolma category in the top quartile for the first three principal components.
56020
Reposted by Stephan Hollander
Per Engzell @pengzell.bsky.social · 12/06/2026
Short story about statistical modeling without causal inference gone badly wrong. Just head on the news that preschoolers play 15 minutes less on rainy days. The reporting stressed that the research was "associational" and therefore couldn't tell us why. Huh? 1/
77826
Reposted by Stephan Hollander
Stefan Feuerriegel @sfeuerriegel.bsky.social · 10/06/2026
📣 New paper @nathumbehav.nature.com: A reporting checklist for large language models in behavioural science www.nature.com/articles/s41...
1178
Reposted by Stephan Hollander
Charlotte Garden @charlottegarden.bsky.social · 06/06/2026
Conference organizers called the cops on attendees for distributing an op-ed (published in a medical journal) that was critical of the Trump administration's health policy. www.nytimes.com/2026/06/05/w...
nytimes.com
Police Remove Diabetes Experts From Conference for Distributing Critique of Trump Administration
12418
Reposted by Stephan Hollander
Paul Hünermund @p-hunermund.com · 01/06/2026
Interesting new paper on AI coding tools and productivity on GitHub. Large gains at the code-writing stage are partly offset by downstream human bottlenecks in code review and deployment. Link: www.nber.org/papers/w35275
75217
Reposted by Stephan Hollander
Andrea Lathrop @cabernet.bsky.social · 26/05/2026
Cross-posting this from Twitter, because I like it: (by @littmath.bsky.social)
23510
Reposted by Stephan Hollander
Peter Tennant @pwgtennant.bsky.social · 19/05/2026
You may have seen this graph in the FT, which claims to show birth rates plunging since the introduction to smartphone. Except it doesn't. Read my short piece with @debscohen.bsky.social on some of the problems with this graph of juicy cherries! unherd.com/newsroom/are...
FT graphic by John Burn-Murdoch entitled 'could digital media be affecting birth rates?'
78429
Reposted by Stephan Hollander
Ethan Mollick @emollick.bsky.social · 19/05/2026
🚨Our paper is out in PNAS: we found classic human persuasion techniques worked on AIs in a "parahuman" way, making them agree to objectionable requests (increasing compliance from 35% to 51%) It worked on a range of major recent LLMs though newer models do resist more www.pnas.org/doi/10.1073/...
0509
Reposted by Stephan Hollander
Rohan @rohandas.net · 11/05/2026
Computational approaches to media narrative analysis either miss nuanced storytelling patterns through coarse-grained analysis, or require domain-specific taxonomies that limit scalability. We show joint event and character modeling can address this gap. Details in our #ACL2026 (Main) paper. 🧵1/10
Paper Title: A Structured Clustering Approach for Inducing Media Narratives

Authors: Rohan Das, Advait Deshmukh, Alexandria Leto, Zohar Naaman, I-Ta Lee, Maria Leonor Pacheco
12411
Reposted by Stephan Hollander
Omar Wasow @owasow.bsky.social · 06/05/2026
“Writing is hard.” Thrilled to share that this simple idea led to a new paper in Political Analysis! Where most text methods focus on content, I test if expression is also effortful action. I find simple measures like character counts reveal attitudes and predict voting. cup.org/4cUmoXi 1/
Text as Behavior
Text as Behavior published in Political Analysis

by Omar Wasow

Abstract

Text analysis typically focuses on content—such as sentiment or topic—but expression is also a form of effortful action. Building on this insight, I propose using simple features of open-ended tasks to study text as behavior. This approach treats expression, such as writing, as cognitively, emotionally and temporally “costly” for subjects but inexpensive for researchers. I show basic statistics like the number of characters can approximate effort and significantly improve estimation of quantities of interest, including candidate choice, the probability of turning out to vote and psychological states about which a subject may not be fully aware. Further, these methods can convert nonresponse into informative data; validate survey instruments; serve as mechanism checks; be hard for a subject to “game”; work across different languages and analogize well to real-world situations. In sum, text as behavior can help address a range of issues related to quantifying attitudes and actions.
11441106
Stephan Hollander @stephanhollander.bsky.social · 09/05/2026
cc @emollick.bsky.social
030
Reposted by Stephan Hollander
Michael R Strain @michaelrstrain.bsky.social · 04/05/2026
🚨New paper. 🧵Our baseline estimate implies that preserving trust in the integrity and quality of official statistics generates economic benefits of about $25 for every $1 spent on the U.S. Bureau of Labor Statistics's budget. cc: @nickbloom.bsky.social, @ericagroshen.bsky.social, #econsky
325790
Reposted by Stephan Hollander
Kyle Lo @ ICML2026 🇰🇷 @kylelo.bsky.social · 01/05/2026
during in Olmo 3 we thought long context is just finding good data nope! model architecture matters & it's hard to recover if mess it up led by @abertsch.bsky.social, we release many pretrain runs w/ small arch changes and show huge long context performance diffs
05811
Reposted by Stephan Hollander
AEA Journals @aeajournals.bsky.social · 28/04/2026
Forthcoming in the AER: "The Effect of Omitted Variables on the Sign of Regression Coefficients" by Matthew A. Masten and Alexandre Poirier.
aeaweb.org
The Effect of Omitted Variables on the Sign of Regression Coefficients
(Forthcoming Article) - We show that, depending on how the impact of omitted variables is measured, it can be substantially easier for omitted variables to flip coefficient signs than to drive them to zero. This behavior occurs with “Oster’s delta” (Oster 2019b), a widely reported robustness measure. Consequently, any time this measure is large—suggesting that omitted variables may be unimportant—a much smaller value reverses the sign of the parameter of interest. We propose a modified measure of robustness to address this concern. We illustrate our results in four empirical applications and two meta-analyses. We implement our methods in the companion Stata module regsensitivity.
0206
Reposted by Stephan Hollander
Christian Zimmermann @czimm-economist.bsky.social · 28/04/2026
To illustrate how much bots are pounding RePEc sites, Google Analytics, which is supposed to weed them out, thinks there have been 3.5M active users from Singapore (population 6M) in the last four weeks. And as many users from Iraq as from Canada. ideas.repec.org #RePEc #EconSky
ideas.repec.org
Economics and Finance Research | IDEAS/RePEc
IDEAS is a central index of economics and finance research, including working papers, articles and software code
3187
Reposted by Stephan Hollander
Paul Goldsmith-Pinkham @paulgp.com · 28/04/2026
My friend is teaching to his Economics Ph.D. students and explaining why they should be interested in topics that are not in their subfield. The friend asked for statements about why learning from other fields is good. Here was my answer: " Five things:
129525
Reposted by Stephan Hollander
John Burn-Murdoch @jburnmurdoch.ft.com · 27/04/2026
The FT @financialtimes.com is hiring a new data journalist to join our US data and visuals team in New York. Great job, great team, great place to work. Apply here 👉 job-boards.eu.greenhouse.io/financialtim...
job-boards.eu.greenhouse.io
US Data Journalist
New York; Washington DC
411372
Reposted by Stephan Hollander
Konrad Burchardi @konradburchardi.bsky.social · 27/04/2026
1. I am looking for a research assistant to work on a project which investigates Sweden's spectacular economic growth during the late 19th century using establishment-level data.
11413
Stephan Hollander @stephanhollander.bsky.social · 26/04/2026
Officially breaking two 🤩
010
Reposted by Stephan Hollander
Petter Törnberg @pettertornberg.com · 23/04/2026
Social media is no longer social. Most of it is passive viewing of videos and pictures from people we've never met. But we're still studying social media like it's 2010. We've entered the post-social media era — and research needs to catch up. osf.io/preprints/so... preprint w/ Richard Rogers
392983945
Reposted by Stephan Hollander
Maria Antoniak @mariaa.bsky.social · 20/04/2026
What are the worst ethical disasters in NLP history? (I'm teaching "ethics of NLP" tomorrow and history is good for teaching this topic.) Most are data breaches/releases (AOL search logs, OKCupid profiles, Finnish therapy records...) but what others? I'll put some other examples in thread --> 1/n
64011
Reposted by Stephan Hollander
Merrilee Proffitt @merrileeiam.bsky.social · 21/04/2026
I'm excited to share that we've made a collection of historic Supreme Court Records and Briefs available via @archive.org I've written a blog post where I go into detail about the importance of this collection. blog.archive.org/2026/04/20/u...
blog.archive.org
U.S. Supreme Court Records and Briefs: The Arguments That Shaped America, Now Freely Available | Internet Archive Blogs
4620234
Reposted by Stephan Hollander
Todd Jones @toddrjones.bsky.social · 17/04/2026
Daily stock returns since 1950.
2144
Reposted by Stephan Hollander
Philine Widmer @phinifa.bsky.social · 14/04/2026
We have started the second day of the MPWZ-CEPR Text-as-Data Workshop (already the 11th edition). Join us here: ethz.zoom.us/j/62143211732 Text-as-data is used across economics now -- from projects on gender norms to mafia networks, and sanctions evasion, to superstar scientists and AI patents 📈📊
122
Reposted by Stephan Hollander
Philine Widmer @phinifa.bsky.social · 13/04/2026
📕 💻 📈 Are you curious about the latest work on text-as-data in economics and beyond? We have just kicked off the 11th MPWZ-CEPR Text-as-Data Workshop! We will "time-travel" with LLMs, track narratives, and much more! 40 papers, one link: ethz.zoom.us/j/62143211732. Program: tinyurl.com/yc2zvy7u
142
Reposted by Stephan Hollander
Institute for Replication @i4replication.bsky.social · 01/04/2026
🧵1/ Our first meta-science paper (with 350+ coauthors) is published today in Nature. It presents one of the largest-ever reproducibility projects in economics & political science. Here’s what we found 👇
216989
Reposted by Stephan Hollander
Soumaya Keynes @soumayakeynes.ft.com · 19/03/2026
found something rather baffling when researching my column this week… I wanted to see if there was any evidence that AI tools were helping economists to make their research more readable. So I analysed the text of NBER working paper abstracts…
12711
Reposted by Stephan Hollander
Maria Antoniak @mariaa.bsky.social · 10/03/2026
I'm lecturing about the "History of NLP" this week. What should I include? Any favorite anecdotes, images, people, methods? Slides, books, papers, or talks for inspiration or grounding? I've been maintaining a small collection here: www.are.na/maria-antoni...
are.na
🗄 history of NLP and the ACL | Are.na
267715
Reposted by Stephan Hollander
Kate Mackenzie @katemac.bsky.social · 08/03/2026
Threading some stuff about oil & oil markets, just basic but hope it helps: 1/ oil markets are what you call “finely balanced”. Supply is usually very very close to demand/consumption. Demand is hard to shift *quickly* in response to supply hiccups. So even small supply changes = big price effects
320484
Stephan Hollander @stephanhollander.bsky.social · 08/03/2026
A lifetime of collecting: the 70,000-volume home library of Bruno Schröder, a mining engineer. A wonderland of books 🤩 www.rarebookhub.com/articles/3355
050
Reposted by Stephan Hollander
Sven-Erik Volberg @volberg.bsky.social · 07/03/2026
Published paper proving that #ChatGPT will always make things up. Not sometimes. Not until the next update. Always. They proved it with math. Even with perfect data and unlimited computing power, AI models will still confidently tell you things that are completely false. arxiv.org/abs/2509.04664
49436
Stephan Hollander @stephanhollander.bsky.social · 06/03/2026
🏷️ @aleximas.bsky.social
062
Reposted by Stephan Hollander
Laura Veldkamp @laura-veldkamp.bsky.social · 05/03/2026
NY is proposing to "Impose liability for damages caused by a chatbot impersonating certain licensed professionals." nysenate.gov/legislation/... How does a chatbot trick you into thinking its a doctor? Senators: If you forgot you were conversing with AI, you need a doctor.
nysenate.gov
NY State Senate Bill 2025-S7263
Imposes liability for damages caused by a chatbot impersonating certain licensed professionals.
032
Reposted by Stephan Hollander
Laura K. Nelson @lauraknelson.bsky.social · 05/03/2026
Ok I'm in a rabbit hole. If you search "how many decisions do we make in a day" the reported number is almost always 35,000, often reported that this is according to "multiple sources". Yet I can't actually find a single source that backs up that number. Anyone know where this number comes from?
5238