Alexander Wuttke
LightNews LightNews
kunkakom.bsky.social
Alexander Wuttke
@kunkakom.bsky.social
3.9K followers 2.2K following 410 posts

Digitalization & Pol Behavior LMU_Muenchen | Democracy and Populist Attitudes | Open Science | 🚵‍♂️🐶

Videos
Political science 63%
Communication & Media Studies 7%
Posts Replies Media Videos

kunkakom.bsky.social
Alexander Wuttke @kunkakom.bsky.social · 30m
The sale pitch for this AI Scientist "Kosmos" as presented in this Podcast just seems like a big HARKing exercise. Yes, just look at the data long enough and you'll find "something".

Has anybody made any experiences with AI scientists models? Recommendations?
pca.st/episode/58aa...
Where Is All the A.I.-Driven Scientific Progress?
pca.st

Reposted by Tim Bale, Alexander Wuttke, Paolo Crosetto , and 4 more

poqjournal.bsky.social
Public Opinion Quarterly - POQ @poqjournal.bsky.social · 8d
Survey experiments have become a popular methodology among social scientists. Has it been effective?

In POQ, Rauf et al. study the efficacy of 100 survey experiments. Their results show that a majority of hypotheses were not supported.

Read now: doi.org/10.1093/poq/...

kunkakom.bsky.social
Alexander Wuttke @kunkakom.bsky.social · 8d
Congratulations!

Reposted by Alexander Wuttke

ianhussey.mmmdata.io
Ian Hussey @ianhussey.mmmdata.io · 8d
I’m vocally skeptical of silicon samples, yet vocally impressed by SurveyBot3000.

The difference: this does not rely on magic beans or assumed omniscience, it is trained and validated against a large corpus of highly relevant data and makes specific predictions with known accuracy and precision.
ruben.the100.ci Ruben C. Arslan @ruben.the100.ci · 8d
Finally, @bjoernhommel.bsky.social's and my paper introducing the SurveyBot3000 is officially out in AMPPS. It's a fine-tuned language model that guesstimates correlations between survey items from text alone. Not perfectly, but useful for search, for example.
journals.sagepub.com/doi/10.1177/...
Fig. 3. Scatter plots of the synthetic and empirical estimates, validation study (Stage 2). Showing N = 30,135 item-pair correlations, N = 257 scale reliabilities, and N = 1,568 scale-pair correlations for (top) the pretrained SBERT model and (bottom) the fine-tuned SurveyBot3000 model. SBERT = all-mpnet-base-v2 model. Fig. 4. Prediction error of the synthetic estimates, validation study (Stage 2). Our prediction model allowed the error term to vary freely according to the predictor, the synthetic estimate. The thin-plate splines show that some synthetic estimates were predictably more accurate. Fig. 5. Accuracy by domain. Accuracy differed across domains. SurveyBot3000 accuracy (colored) was always higher than SBERT accuracy (gray). Results were largely consistent whether accuracy of items was tested (left, circle) within domains or (right, cross) across domains. Fig. 1. Multistep training procedure for the SurveyBot3000, which produces synthetic estimates of interitem correlations. (a) Pretraining base model (SBERT). (b) Fine-tuning SurveyBot3000. (c) Validation. SBERT = all-mpnet-base-v2 model.

Reposted by Alexander Wuttke

ruben.the100.ci
Ruben C. Arslan @ruben.the100.ci · 8d
We took pains to avoid overfitting. In addition to the standard training/test/holdout divide, we also ran a registered report follow up and locked down the predictions before we collected data. Accuracy was indeed a tad attenuated r=.71 -> .59.

Reposted by Alexander Wuttke, Cameron Brick, Ian Hussey , and 3 more

ruben.the100.ci
Ruben C. Arslan @ruben.the100.ci · 8d
Finally, @bjoernhommel.bsky.social's and my paper introducing the SurveyBot3000 is officially out in AMPPS. It's a fine-tuned language model that guesstimates correlations between survey items from text alone. Not perfectly, but useful for search, for example.
journals.sagepub.com/doi/10.1177/...
Fig. 3. Scatter plots of the synthetic and empirical estimates, validation study (Stage 2). Showing N = 30,135 item-pair correlations, N = 257 scale reliabilities, and N = 1,568 scale-pair correlations for (top) the pretrained SBERT model and (bottom) the fine-tuned SurveyBot3000 model. SBERT = all-mpnet-base-v2 model. Fig. 4. Prediction error of the synthetic estimates, validation study (Stage 2). Our prediction model allowed the error term to vary freely according to the predictor, the synthetic estimate. The thin-plate splines show that some synthetic estimates were predictably more accurate. Fig. 5. Accuracy by domain. Accuracy differed across domains. SurveyBot3000 accuracy (colored) was always higher than SBERT accuracy (gray). Results were largely consistent whether accuracy of items was tested (left, circle) within domains or (right, cross) across domains. Fig. 1. Multistep training procedure for the SurveyBot3000, which produces synthetic estimates of interitem correlations. (a) Pretraining base model (SBERT). (b) Fine-tuning SurveyBot3000. (c) Validation. SBERT = all-mpnet-base-v2 model.

kunkakom.bsky.social
Alexander Wuttke @kunkakom.bsky.social · 10d
Same.
And users start to recognize when others change their profile pictures...

kunkakom.bsky.social
Alexander Wuttke @kunkakom.bsky.social · 10d
Is it just me or has bluesky started to feel like good old academic twitter recently?

kunkakom.bsky.social
Alexander Wuttke @kunkakom.bsky.social · 10d
Australia bans social media for the youth.
Any plans that
@jonathanhaidt.bsky.social and his critics get together for an adversarial collaboration to pre-register success/failure criteria on a specific diff-in-diff design to exploit this natural experiment?
pca.st/episode/0155...
Australia Kicks Kids Off Social Media + Is the A.I. Water Issue Fake? + Hard Fork Wrapped
pca.st

Reposted by Alexander Wuttke

deevybee.bsky.social
Dorothy Bishop @deevybee.bsky.social · 14d
i guess the expertise of the authors is limited to experience of being criticised
ianhussey.mmmdata.io Ian Hussey @ianhussey.mmmdata.io · 14d
Critique argues post-publication peer review "often focuses on minor details" and "risks damaging trust in both the research itself and the peer review process" found to contain:

- Hallucinated references
- Undisclosed COIs (EiC is author)
- 8/9 authors have retractions or related scandals
An expert criticism on post-publication peer review platforms: the case of pubpeer - DARU Journal of Pharmaceutical Sciences
While traditional peer review offers advantages in academic publishing, it is often hampered by significant weaknesses, leading to frustration among many authors. Scientific discoveries after publicat...
link.springer.com

Reposted by Dorothy Bishop, Alexander Wuttke, Flávio Azevedo

ianhussey.mmmdata.io
Ian Hussey @ianhussey.mmmdata.io · 14d
Critique argues post-publication peer review "often focuses on minor details" and "risks damaging trust in both the research itself and the peer review process" found to contain:

- Hallucinated references
- Undisclosed COIs (EiC is author)
- 8/9 authors have retractions or related scandals
An expert criticism on post-publication peer review platforms: the case of pubpeer - DARU Journal of Pharmaceutical Sciences
While traditional peer review offers advantages in academic publishing, it is often hampered by significant weaknesses, leading to frustration among many authors. Scientific discoveries after publicat...
link.springer.com

Reposted by Alexander Wuttke, John Holbein, Kai Gehring , and 1 more

paulgp.com
Paul Goldsmith-Pinkham @paulgp.com · 14d
I was also bored and pulled up the replication package.

The results are quite sensitive in aggregate to needing the controls. Here's the replication of Figure 3, including the case w/o controls:

Reposted by Alistair Munro, Alexander Wuttke, John Holbein , and 1 more

chanret.bsky.social
Chris Hanretty @chanret.bsky.social · 14d
Thanks to @johnholbein1.bsky.social I learned about this paper on rent control in Berlin.

Because I was marking, I immediately downloaded the replication materials.

bsky.app/profile/john...
johnholbein1.bsky.social John Holbein @johnholbein1.bsky.social · 15d
Um, ok...

This paper forthcoming at the JOP provides evidence that rent control in Germany actually made tenants MASSIVELY *less* NIMBY.

This result was in the opposite direction of the authors' pre-registered expectations.

And the effect sizes are, truly, massive.

Reposted by Alexander Wuttke, Martin Kreidl, Timothy Graham , and 4 more

css-lmu.bsky.social
CSS Lab, LMU Munich @css-lmu.bsky.social · 15d
🚀 We’re hiring a Postdoc!

Our group is looking for a Postdoc to join the team working on computational comm research. If you’re excited about automated content analysis, large text & social-media data, open science, this might be for you.

💡 Sounds like you or someone you know? Please share/boost!

Reposted by Richard S.J. Tol, Jörg Peters, Alexander Wuttke , and 2 more

i4replication.bsky.social
Institute for Replication @i4replication.bsky.social · 16d
We're thrilled to open registration for our 1st 2026 Replication Games. The event will be at the University of Zurich on January 19th.

Psych, public health, pol sci and econ studies will be reproduced! Register here: www.surveymonkey.ca/r/Replicatio...

Reposted by Alessandro Nai

kunkakom.bsky.social
Alexander Wuttke @kunkakom.bsky.social · 16d
Grateful for the opportunity to contribute—by name—to what must be the most massive collaborative encyclopedia project of all time, next to Wikipedia. All credit goes to @annisch.bsky.social who did the heavy lifting for the overview article on Populist Attitudes.

osf.io/preprints/so...

Reposted by Joanna Bryson, Alexander Wuttke

lorenzspreen.bsky.social
Philipp Lorenz-Spreen @lorenzspreen.bsky.social · 16d
Christmas came early this year! Very happy to see our paper out in Science Advances. Led by @lfoswaldo.bsky.social, we ran a unique collective field-experiment on Reddit, to better understand who is participating in online debates and why.

Paper: www.science.org/doi/10.1126/...

And more below 👇

Reposted by Alexander Wuttke

ajungherr.bsky.social
Andreas Jungherr @ajungherr.bsky.social · 17d
📢 New paper out!
What do people want from AI systems? How should outputs be adjusted? And how do views differ between countries?
@adrauc.bsky.social and I explore this for @socialmedia-soc.bsky.social in Public Opinion on the Politics of AI Alignment.

journals.sagepub.com/doi/10.1177/...

kunkakom.bsky.social
Alexander Wuttke @kunkakom.bsky.social · 17d
Oh, I see.
you're right. Journals retracting the article is, of course, not the only and perhaps not even the most common way.
Good point

kunkakom.bsky.social
Alexander Wuttke @kunkakom.bsky.social · 17d
I like this idea.

Let's include retraction in university rankings
(And perhaps successful reproductions)
ianhussey.mmmdata.io Ian Hussey @ianhussey.mmmdata.io · 17d
Achal Agrawal is on Nature’s list of 10 people who shaped science in 2025. His work helped change India’s university rankings system to include a penalty for large number of retractions.

www.nature.com/articles/d41...
This science sleuth revealed a retraction crisis at Indian universities
Achal Agrawal is part of Nature’s 10, a list of people who shaped science in 2025.
www.nature.com

Reposted by Dorothy Bishop, Alexander Wuttke

ianhussey.mmmdata.io
Ian Hussey @ianhussey.mmmdata.io · 17d
Achal Agrawal is on Nature’s list of 10 people who shaped science in 2025. His work helped change India’s university rankings system to include a penalty for large number of retractions.

www.nature.com/articles/d41...
This science sleuth revealed a retraction crisis at Indian universities
Achal Agrawal is part of Nature’s 10, a list of people who shaped science in 2025.
www.nature.com

kunkakom.bsky.social
Alexander Wuttke @kunkakom.bsky.social · 18d
wow. So nice to hear. Congratulations and good luck!

Reposted by Gijs Schumacher, Alexander Wuttke

mrooduijn.bsky.social
Matthijs Rooduijn @mrooduijn.bsky.social · 18d
With my project IMMERSE I investigate whether immersive narrative interventions (storytelling, live theatre, video games and virtual reality) can strengthen support for liberal democracy by engaging people more deeply than conventional communication. See: www.uva.nl/en/content/n... 2/2
Nine UvA researchers receive ERC Consolidator Grants
The European Research Council (ERC) has awarded Consolidator Grants to nine UvA researchers. The laureates are: Thijs Bol, Jaron Harambam, Eirini Karyotaki, Patty Leijten, Dora Matzke, Matthijs Roodui...
www.uva.nl

kunkakom.bsky.social
Alexander Wuttke @kunkakom.bsky.social · 18d
Perhaps it should include a slide on the proliferation of causal designs in polisci in the recent 2 decades as documented by @carotorreblanca.bsky.social
et Al
osf.io/preprints/so...
OSF
osf.io

kunkakom.bsky.social
Alexander Wuttke @kunkakom.bsky.social · 18d
I'm grateful for critical views and hints

kunkakom.bsky.social
Alexander Wuttke @kunkakom.bsky.social · 18d
I played around with NotebookLM and asked it to build a short explainer based on "Causal Empiricism", a @nickchk.com chapter and different prompts. That’s what the video above is based on.

kunkakom.bsky.social
Alexander Wuttke @kunkakom.bsky.social · 18d
I’d still love to have one resource that introduces the logic of (a) controlled and (b) natural experiments and (c) directly contrasts these designs with the “just run a regression” approach. Do you know readings / videos / other material that do all of this well?

kunkakom.bsky.social
Alexander Wuttke @kunkakom.bsky.social · 18d
Today I noticed that my BA students struggled with it. My fault: I used it to introduce our causal design block instead of placing it at the end. The text assumes that readers already know the main designs.

kunkakom.bsky.social
Alexander Wuttke @kunkakom.bsky.social · 18d
A while ago I asked here which text to assign on this. Many of you (thanks!) pointed me to the @cdsamii.bsky.social piece and it has become one of my absolute favorite readings in 2025.
www.journals.uchicago.edu/doi/abs/10.1...
Causal Empiricism in Quantitative Research | The Journal of Politics: Vol 78, No 3
Quantitative analysis of causal effects in political science has trended toward the adoption of “causal empiricist” approaches. Such approaches place heavy emphasis on causal identification through ex...
www.journals.uchicago.edu

Reposted by Cyrus Samii, David Darmofal

kunkakom.bsky.social
Alexander Wuttke @kunkakom.bsky.social · 18d
How to explain the “credibility revolution” in a few minutes?

"I" "made" a short video that tries to show why clever (natural) experiments and research design beat pure statistical adjustment for causal claims.

I am genuinely curious what methods people think:
youtu.be/Fv14ktwA31Q?...
The Causal Revolution: Why Research Design trumps (regression) models for causal claims
YouTube video by Alexander Wuttke
youtu.be