Johannes Breuer's Avatar

Johannes Breuer

@johannesbreuer.com

Professor of Digital Social Science @unidue.bsky.social and head of the team Research Data & Methods @cais-research.bsky.social Interested in digital traces | computational social science | reproducibility | open science | #rstats www.johannesbreuer.com

2,670
Followers
2,155
Following
258
Posts
25.09.2023
Joined
Posts Following

Latest posts by Johannes Breuer @johannesbreuer.com

Post image
13.03.2026 20:46 ๐Ÿ‘ 27 ๐Ÿ” 4 ๐Ÿ’ฌ 1 ๐Ÿ“Œ 0
Preview
Episode 77: Miscitatio | Nullius in Verba In this episode, we discuss the problem of miscitation. How often are citations to the scientific literature outright misleading? Do we really need to spell out that peopleย  are supposed to read whatโ€ฆ

New Nullius In Verba episode on miscitations:
nulliusinverba.podbean.com/e/ep-77-misc...
How often are citations to the scientific literature outright misleading? Do we really need to spell out that people are supposed to read what they cite? And should we review citations as they do in law?

13.03.2026 21:06 ๐Ÿ‘ 10 ๐Ÿ” 2 ๐Ÿ’ฌ 0 ๐Ÿ“Œ 0
paper title: LLM-Assisted Replication as Scientific Infrastructure. paper authors:  So Kubota, Hiromu Yakura, Sho Yamada, Yuki Nakamura, Tobias Werner, Samuel Coavoux

paper title: LLM-Assisted Replication as Scientific Infrastructure. paper authors: So Kubota, Hiromu Yakura, Sho Yamada, Yuki Nakamura, Tobias Werner, Samuel Coavoux

Abstract: Large language models (LLMs) are rapidly accelerating scientific production, from literature synthesis to automated analysis. Yet this expansion risks creating a verification gap, in which the volume of scientific claims outpaces the communityโ€™s capacity to check their reproducibility. We argue that the same LLM capabilities driving scientific output can be redirected toward scalable verification. As a demonstration, we share insights and lessons from our attempt to reproduce the core results of popular classical papers across disciplines, which yielded not only successful cases but also failures due to underspecified methodological details. This means that automated replication does not adjudicate scientific truth, but it localizes discrepancies and documentation gaps, lowering the cost of computational reproducibility checks. We thereby propose embedding LLM-assisted replication across the research lifecycle, from pre-submission quality check, journal-integrated verification, post-publication audits, to forensic reconstruction of legacy studies. To prevent misuse and preserve trust, we call for transparent standards and community governance. If institutionalized responsibly, AI can serve not only to generate science, but to scale its self-correction.

Abstract: Large language models (LLMs) are rapidly accelerating scientific production, from literature synthesis to automated analysis. Yet this expansion risks creating a verification gap, in which the volume of scientific claims outpaces the communityโ€™s capacity to check their reproducibility. We argue that the same LLM capabilities driving scientific output can be redirected toward scalable verification. As a demonstration, we share insights and lessons from our attempt to reproduce the core results of popular classical papers across disciplines, which yielded not only successful cases but also failures due to underspecified methodological details. This means that automated replication does not adjudicate scientific truth, but it localizes discrepancies and documentation gaps, lowering the cost of computational reproducibility checks. We thereby propose embedding LLM-assisted replication across the research lifecycle, from pre-submission quality check, journal-integrated verification, post-publication audits, to forensic reconstruction of legacy studies. To prevent misuse and preserve trust, we call for transparent standards and community governance. If institutionalized responsibly, AI can serve not only to generate science, but to scale its self-correction.

Post image Post image

"LLM-Assisted Replication as Scientific Infrastructure", V2 of our preprint about how to use LLM to make better rather than more papers and assess the extent of the replication crisis in sociology osf.io/preprints/so...

13.03.2026 21:00 ๐Ÿ‘ 3 ๐Ÿ” 2 ๐Ÿ’ฌ 0 ๐Ÿ“Œ 0

Join us next week for an exciting grand opening w/ Max in the TADA.cool spring ๐ŸŒฑ speaker series on LLMs in survey research โฌ‡๏ธ

13.03.2026 20:10 ๐Ÿ‘ 1 ๐Ÿ” 3 ๐Ÿ’ฌ 0 ๐Ÿ“Œ 0
Post image

I build a feed for european grant funding that's updated daily and (IMO) far more transparent then the official EU fundign website: michelnivard.github.io/eu-grants-fe... (inspired by the NIH equivalent @sashagusevposts.bsky.social build yesterday)

13.03.2026 11:19 ๐Ÿ‘ 42 ๐Ÿ” 20 ๐Ÿ’ฌ 3 ๐Ÿ“Œ 1
Preview
Dr. Marco Wรคhner erhรคlt Projektfรถrderung von DGOF - CAIS Dr. Marco Wรคhner, wissenschaftlicher Mitarbeiter im Team Research Data & Methods am CAIS, wird mit einer Projektfรถrderung der Deutschen Gesellschaft fรผr Online-Forschung (DGOF) ausgezeichnet. In seine...

Wir freuen uns mit Marco Wรคhner, dass sein Projekt von der DGOF gefรถrdert wird!๐Ÿฅณ
Er untersucht darin, welche Faktoren Bรผrger:innen motivieren, an Online-Konsultationen zu lokaler Mobilitรคt und Infrastrukturplanung teilzunehmen.
โ„น๏ธ www.cais-research.de/news/dr-marc...
@marco-waehner.bsky.social

12.03.2026 07:47 ๐Ÿ‘ 8 ๐Ÿ” 2 ๐Ÿ’ฌ 0 ๐Ÿ“Œ 0
Visual zur Podcastfolge mit dem Titel "Willkommen im Gruselgraben: Kรผnstliche Avatare und echtes Vertrauen in der Wissenschaftskommunikation" und drei Icons, die Avatar, Graben und Vertrauen in Wissenschaftskommunikation symbolisieren

Visual zur Podcastfolge mit dem Titel "Willkommen im Gruselgraben: Kรผnstliche Avatare und echtes Vertrauen in der Wissenschaftskommunikation" und drei Icons, die Avatar, Graben und Vertrauen in Wissenschaftskommunikation symbolisieren

๐ŸŽ™๏ธFast menschlich โ€“ und genau deshalb unheimlich?
Im #CAISzeit-Podcast spricht Jasmin Baake mit Dr. Matthias Begenat รผber KI-Avatare in der Wissenschaftskommunikation: Wann schaffen sie Vertrauen โ€“ und wann landen wir im Uncanny Valley?
www.cais-research.de/news/willkom...
@herr-bege.bsky.social

12.03.2026 08:16 ๐Ÿ‘ 3 ๐Ÿ” 2 ๐Ÿ’ฌ 0 ๐Ÿ“Œ 0
Preview
ChatGPT Edu feature reveals researchersโ€™ project metadata across universities (exclusive) A configuration in Codex Cloud Environments lets thousands of colleagues see repository names and activity linked to ChatGPT accounts.

If you are a researcher using ChatGPT Edu metadata about your research has been leaking www.fastcompany.com/91507219/cha...

12.03.2026 11:18 ๐Ÿ‘ 23 ๐Ÿ” 13 ๐Ÿ’ฌ 1 ๐Ÿ“Œ 3
Announcements | Journal of Open Psychology Data

If anyone is interested in a journal leadership position, there is a call for editor-in-chief for the Journal of Open Psychology Data: openpsychologydata.metajnl.com/announcements. I'm on the board, so I'm happy to chat to anyone or put your in touch with others!

11.03.2026 21:48 ๐Ÿ‘ 8 ๐Ÿ” 6 ๐Ÿ’ฌ 0 ๐Ÿ“Œ 0
Post image

Save the date: the ODISSEI Conference on 24 November celebrates a decade of advancing national research infrastructure for the social sciences.

Join computational social scientists to discuss the needs of digital and computational social science research.

Learn more: edu.nl/7cfnb

11.03.2026 08:44 ๐Ÿ‘ 1 ๐Ÿ” 1 ๐Ÿ’ฌ 0 ๐Ÿ“Œ 0
Video thumbnail

๐Ÿฅ๐Ÿฅ๐Ÿฅ Newly out from us today in Science Advances: โ€œBiased AI Writing Assistants Shift Usersโ€™ Attitudes on Societal Issuesโ€.

Large Language Models are providing users with autocomplete writing suggestions on many platforms. Could these suggestions shift usersโ€™ own attitudes? (spoiler: YES) (1/7)

11.03.2026 19:02 ๐Ÿ‘ 144 ๐Ÿ” 79 ๐Ÿ’ฌ 4 ๐Ÿ“Œ 12
Scientific career at GESIS
Scientific career at GESIS At GESIS, researchers work in various phases of their scientific careers: as student assistants, doctoral candidates, postdocs, senior researchers, team leaders, and department heads. The institute's goal is to provide its scientific staff with the best possible support in their various career phase

#jobs #stellenangebote #GESISjobs #jobfairy
Pursue your scientific career at GESIS!
youtu.be/zMY2zqYFIMg

Apply now for a job as a #ResearchAssociate (#PhD Candidate) Cross-platform Social Media Mining
(Salary group 13 TV-L, working time 100 %, limited to 36 months)

gesis.jobs.personio....

06.03.2026 13:01 ๐Ÿ‘ 1 ๐Ÿ” 2 ๐Ÿ’ฌ 0 ๐Ÿ“Œ 0
Post image

Introducing the German General Social Survey Personas (GGSSP) โ€“ a #persona collection derived from the German General Social Survey (#GGSS #ALLBUS).

This data release provides plug-and-play persona descriptions to steer the generations of LLMs towards individual members of the German population.

10.03.2026 08:00 ๐Ÿ‘ 16 ๐Ÿ” 9 ๐Ÿ’ฌ 2 ๐Ÿ“Œ 1
Arbeiten bei GESIS
Arbeiten bei GESIS Profitieren Sie von vielen Vorteilen! Wir stellen attraktive Arbeitsplรคtze in Kรถln oder Mannheim bereit und bieten Ihnen groรŸzรผgige Homeoffice-Regelungen. Bei uns wird die Vereinbarkeit von Beruf und Familie groรŸgeschrieben und unsere Mitarbeitenden profitieren von unserem betrieblichen Gesundheits

#jobs #stellenangebote #GESISjobs #jobfairy
Arbeiten Sie bei GESIS und profitieren Sie von vielen Vorteilen.
youtu.be/t1KcoQxuvmQ...

Wir suchen fรผr #Mannheim eine/n Wissenschaftliche*n Mitarbeiter*in (#Doktorand*in) #GESISPanel
(Entgeltgruppe 13 TV-L, AZ 75 %, 4 Jahre)

gesis.jobs.personio....

11.03.2026 08:00 ๐Ÿ‘ 2 ๐Ÿ” 3 ๐Ÿ’ฌ 0 ๐Ÿ“Œ 0

Interesting PhD opportunity at GESIS focusing on cross-platform social media mining and group-specific disinformation. Great fit for computational social science researchers interested in communication and AI ethics applications.

#AcademicJobs #ComputationalSocialScience

gesis.jobs.personio.de...

11.03.2026 11:07 ๐Ÿ‘ 0 ๐Ÿ” 2 ๐Ÿ’ฌ 0 ๐Ÿ“Œ 0
The Academic Wheel of Privilege showing the 24 socio-cultural identities. The 24 socio-cultural identity types span six sectors: health and wellbeing, society, culture and communication, gender and sexuality, education and career, living arrangements and lastly childhood and development. These identity types are shown as circles connected to three concentric rings (outer, middle and inner) of โ€œidentityโ€ circles with increasing privilege as you go towards the centre.

The Academic Wheel of Privilege showing the 24 socio-cultural identities. The 24 socio-cultural identity types span six sectors: health and wellbeing, society, culture and communication, gender and sexuality, education and career, living arrangements and lastly childhood and development. These identity types are shown as circles connected to three concentric rings (outer, middle and inner) of โ€œidentityโ€ circles with increasing privilege as you go towards the centre.

Out now!

The Academic Wheel of Privilege ๐ŸŽก

We developed a framework & app to guide authorship teams in making equitable and thoughtful authorship decisions.

@saralilplants.bsky.social, @justinsulik.bsky.social, Bethan Iley, Mahmoud Elsherif, @flavioazevedo.bsky.social

๐Ÿ”— osf.io/preprints/me...

09.03.2026 17:15 ๐Ÿ‘ 51 ๐Ÿ” 28 ๐Ÿ’ฌ 1 ๐Ÿ“Œ 5
Preview
Lunchtime Talk: Enhancing Replicability in Computational Communication Science โ€“ A Proactive Approach - CAIS The emergence of the replication crisis has placed the integrity of research across the social and behavioral sciences under increased scrutiny. While awareness among the community has grown in recent...

Join us for this weekโ€™s lunchtime talk on
"Enhancing Replicability in Computational Communication Science" with Prof. Dr. Johannes Breuer & Xinyue Zhao
๐Ÿ“… 11 March 2026 | 1:30 โ€“ 2:30 PM
๐Ÿ“ at CAIS + Online (contact: kolleg@cais-research.de)
www.cais-research.de/en/event/lun...
@johannesbreuer.com

10.03.2026 10:44 ๐Ÿ‘ 2 ๐Ÿ” 1 ๐Ÿ’ฌ 0 ๐Ÿ“Œ 0
OSF

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/ps...
๐Ÿงต

07.03.2026 20:22 ๐Ÿ‘ 111 ๐Ÿ” 51 ๐Ÿ’ฌ 4 ๐Ÿ“Œ 4
A digital CAPTCHA verification window titled "Select all squares with PIPES" against a plain white background. The window contains a 3ร—3 grid of numbered squares, mixing literal hardware, smoking pipes, and programming syntax.

A digital CAPTCHA verification window titled "Select all squares with PIPES" against a plain white background. The window contains a 3ร—3 grid of numbered squares, mixing literal hardware, smoking pipes, and programming syntax.

These captchas just keep getting harder #rstats

05.03.2026 15:58 ๐Ÿ‘ 620 ๐Ÿ” 142 ๐Ÿ’ฌ 8 ๐Ÿ“Œ 2

โ€ผ๏ธEs ist soweit! Die Wahl des #RatSWD lรคuft vom 2.-29.3., die Einladung zur online-Wahl ging heute raus.

Fรผr die #Politikwissenschaft kandidieren
@ackermannk.bsky.social,
@wahlforschung.thorstenfaas.de und
Lars Rinsdorf

๐Ÿ‘‰ www.konsortswd.de/ueber-uns/ra...

#PoWi #DVPW #DGfP #DGPuK

02.03.2026 14:19 ๐Ÿ‘ 7 ๐Ÿ” 7 ๐Ÿ’ฌ 0 ๐Ÿ“Œ 0

Congratulations, @philippdarius.bsky.social ! And all the best for the new position!

05.03.2026 21:24 ๐Ÿ‘ 2 ๐Ÿ” 0 ๐Ÿ’ฌ 1 ๐Ÿ“Œ 0
Preview
The effects of Facebook and Instagram political advertisements on the 2020 US election Nature Human Behaviour, Published online: 05 March 2026; doi:10.1038/s41562-025-02329-9To study the effects of political advertising, we conducted a field experiment with over 60,000 participants. Removing political advertisements from the Facebook and Instagram feeds of randomly selected participants before the 2020 US election did not have a detectable effect on political knowledge, polarization, turnout or political participation.

The effects of Facebook and Instagram political advertisements on the 2020 US election

05.03.2026 12:16 ๐Ÿ‘ 15 ๐Ÿ” 11 ๐Ÿ’ฌ 0 ๐Ÿ“Œ 7

This an enormous null effect. Social media ads did nothing in terms of voting.

05.03.2026 21:20 ๐Ÿ‘ 10 ๐Ÿ” 1 ๐Ÿ’ฌ 0 ๐Ÿ“Œ 0
Preview
BahnBet โ€” Bet on German Train Delays Predict Deutsche Bahn delays and win. Real-time prediction markets on ICE, IC & EC train delays.

What I thought was a joke is real: you can now bet on German train delays. bahn.bet

04.03.2026 10:29 ๐Ÿ‘ 197 ๐Ÿ” 77 ๐Ÿ’ฌ 15 ๐Ÿ“Œ 33

โฐ Just a friendly reminder that the application deadline for the "PhD position on Human-Computer Interaction, Artificial Intelligence, and Disinformation" ends on March 1, 2026.

If you know someone in your network who would be a strong fit for this position, please share this opportunity with them.

25.02.2026 07:12 ๐Ÿ‘ 2 ๐Ÿ” 2 ๐Ÿ’ฌ 0 ๐Ÿ“Œ 0
Post image

#MeetTheExperts #mte #KODAQS #Toolbox #TES
Don't miss the talk by our expert Leon Frรถhling about the Total Error Sheets for Datasets (TES-D) which offer a structured, template-based tool for documenting datasets collected from online platforms.
Register now: us02web.zoom.us/meet...

03.03.2026 08:00 ๐Ÿ‘ 1 ๐Ÿ” 2 ๐Ÿ’ฌ 0 ๐Ÿ“Œ 0
Scientific career at GESIS
Scientific career at GESIS At GESIS, researchers work in various phases of their scientific careers: as student assistants, doctoral candidates, postdocs, senior researchers, team leaders, and department heads. The institute's goal is to provide its scientific staff with the best possible support in their various career phase

#jobs #stellenangebote #GESISjobs #ComputationalSocialScience
Pursue your scientific career at GESIS!
youtu.be/zMY2zqYFIMg

We offer a job as a Doctoral Researcher in Computational Social Science
(Salary group 13 TV-L, working time 100 %, limited until 31.03.2029):

gesis.jobs.personio....

03.03.2026 13:00 ๐Ÿ‘ 2 ๐Ÿ” 6 ๐Ÿ’ฌ 0 ๐Ÿ“Œ 0

Sharing a great opportunity for those in our field: GESIS is hiring a Doctoral Researcher in Computational Social Science (full-time, funded through March 2029). Excellent place to develop CSS skills in a leading research environment.

#AcademicJobs #ComputationalSocialScience

youtu.be/zMY2zqYFIMg

04.03.2026 11:25 ๐Ÿ‘ 1 ๐Ÿ” 2 ๐Ÿ’ฌ 0 ๐Ÿ“Œ 0
Preview
PhD Position: Factual Belief Polarisation and its Attitudinal Consequences | Radboud University Do you want to work as a PhD at the Nijmegen School of Management? Check our vacancy!

๐Ÿšจ Job alert! ๐Ÿšจ

We are looking for a PhD candidate for my Vidi project on factual belief polarization and its attitudinal consequences.

The position will be at the Department of Political Science of Radboud University (Netherlands). Check out the full job profile below!

www.ru.nl/en/working-a...

03.03.2026 11:16 ๐Ÿ‘ 14 ๐Ÿ” 21 ๐Ÿ’ฌ 0 ๐Ÿ“Œ 0
Post image

Want to learn about computational social science *for free* and identify new research partners across academic fields? Apply to one of the 2026 Summer Institutes in Computational Social Science (described in yellow in the attached map) here: sicss.io/locations

03.03.2026 15:01 ๐Ÿ‘ 32 ๐Ÿ” 31 ๐Ÿ’ฌ 0 ๐Ÿ“Œ 0