Hjalmar Bang Carlsen

New working paper with Tom Robinson, Laura Fung and Caroline Roberts. We use an LLM to classify occupations in surveys in real time, probing 'intelligently' when more info is needed. Results show big reductions in cost, time and respondent burden.

osf.io/preprints/so...

Looks like a super useful application.

The Human Toll The US is undergoing rapid social, economic, and policy change. If you've been impacted, we want to hear your story! To participate, visit: https://tinyurl.com/thehumantoll Study sponsored under University of Wisconsin-Madison IRB Protocol #2025-1096. PI: Dr. Jessica Calarco Participants must be at least 18 years of age and reside in the US [image of a person with their face in their hands]

If you want to share your story--or share what you'd say to the people responsible for the current changes in the US--you can do so at the link below. And please share widely! It's essential that we document the human toll.

tinyurl.com/thehumantoll

Get the latest on social science methods and AI

The "can you have" of this world.

My department at Leuphana University offers 6 fully funded PhD scholarships (incl. research funding) for three years.

If you are interested in studying democratic resilience (particularly party competition and elections 🤓) and you have a strong methodolgical background: Apply!

shorturl.at/kcTbG

Assistant Professor - Madison, Wisconsin, United States Current Employees: If you are currently employed at any of the Universities of Wisconsin, log in to Workday to apply through the internal application process.Job Category:FacultyEmployment Type:Regula...

We're hiring! Please share with your networks!

The Department of Sociology at the University of Wisconsin-Madison invites applications for a tenure-track position at the rank of assistant professor in the field of demography of health and aging beginning in August 2026
jobs.wisc.edu/jobs/assista...

Super important agenda paper on the use of AI in survey research. Good to see that AI as interviewer is getting the attention it deserves(most have focus on AI as respondent).

The Stanford GSB is hiring in macro & micro OB! If you do sociology, psychology, OB, strategy, management, computational social science, please apply.

On the macro side, mid-career folks are especially invited to apply.

See here👇
www.gsb.stanford.edu/jobs/faculty...

Vienna is pretty high on my list of liveable cities and so it should be on yours... consider applying

Want to access the slides and the notebook for this session on LLM for Content Analysis in the Social Sciences? Here they are! www.css.cnrs.fr/llm-power-to....

Information about Text classification; Information Extraction, Text Clustering with encoders and decoder models.

We draw on prior efforts of @felixchopra.bsky.social,
@kunkakom.bsky.social, @ihaal.bsky.social, Friedrich Geiecke and Xavier Jaravel.

In the context of #IC2S2 2025:

I would like to hear your thoughts on the future of Computational Social Science.

What do you see as the biggest opportunities and challenges ahead?

It’s a short, AI-led interview.
Interested? Start your interview here:

www.ainterviewer.dk/interview?id...

📢 I’ll be presenting our new work on AI-led vs. human-led chat-based interviews at #IC2S2 in Norrköping this week!

Curious about AI-interviewing, large-N qual data, building your own chatbot for research, or collaborating? Please reach out(DM is open)

For more see thread.

Preprint: osf.io/k7f89

Thanks to my co-authors Semra, Tobias and @andbjn.bsky.social and to the @villumfonden.bsky.social foundation and @cphsodas.bsky.social who have both founded the AI-interview project.

Of course, there are big challenges to this kind of experimental comparison 🧵

1) Data quality in qual interviews isn’t standardized
2) It’s rarely operationalized as quantitative measures
3) Chat-based interviews ≠ in-depth, face-to-face interviews
4) Small N

Important limitations to keep in mind.

📄 Read the preprint: osf.io/k7f89

If you're working on or curious about AI-interview methods, feel free to reach out (DMs open)!

This is part of a broader project developing the AI-interviewer, led together with @annarogers.bsky.social

Why this matters 🧵

1) AI-interviewers can yield reasonable data quality, supporting their promise for large-scale qualitative research.

2) As opposed to other AI-interview studies, we used a locally hosted, open-weight model—boosting data security, transparency & reproducibility.

Key results 🧵:
🧠 Human-led: longer answers per question
🕐 AI-led: longer interviews overall (faster question delivery)
📋 No dif in response quality (specificity + relevance)

LLMs do pretty well—
But responses are short compared typical qualitative interviews.

We ran an experiment with 40 students 🧵.
Participants were randomly assigned to be interviewed—via text chat—by either:

🤖 a locally hosted AI chatbot
👩‍🏫 or a trained human interviewer

Same interview guide. Same interface.
Different interviewers formulating different probes.

What did we find?

🚨 New preprint!
"Comparing AI-led to human-led chat-based interviews"
Can LLMs collect interview data similar to human RAs?

Turns out… yes (mostly).

Note: these were text interviews—not directly comparable to in-depth, face-to-face qualitative interviews.

Led by Semra Guven & @gaardhus.dk
🧵👇

I do not hope that citizens are using chatgpt for learning about politics, at least not DK and other smaller countries. It places Enhedslisten together with Danish People party and Ny Borgerlig on the issue of immigration.

Do Large Language Models Solve the Problems of Agent-Based Modeling? A Critical Review of Generative Social Simulations Recent advancements in AI have reinvigorated Agent-Based Models (ABMs), as the integration of Large Language Models (LLMs) has led to the emergence of ``generative ABMs'' as a novel approach to simula...

Large Language Model-based social simulation has emerged as an exciting new research method.

But do LLMs actually resolve the problems that have historically limited use of Agent-Based Models? What do they bring?

We review the literature to find out!

with Maik Laaroij
arxiv.org/abs/2504.03274

Postdoc position at MIT IDSS focused on collective judgment & decision-making in human–AI hybrid networks. Please apply or share!
academicjobsonline.org/ajo/jobs/29859

UCL – University College London UCL is consistently ranked as one of the top ten universities in the world (QS World University Rankings 2010-2022) and is No.2 in the UK for research power (Research Excellence Framework 2021).

Hiring again! Lecturer in Sociology. We're looking for someone strongly anchored in sociology through high-profile contributions, and with the ability to appeal to students and colleagues of various methodological orientations.
www.ucl.ac.uk/work-at-ucl/...

🔬🧪🧬 SCIENTISTS! we are collecting stories from folks whose research has been disrupted by the Trump admin. Whether you’re a grad student, postdoc, faculty, we want to hear from you! ty to everyone who has already shared.

my signal: caitlingilbert.24

thewashingtonpost.formstack.com/forms/scient...

Join our group in Leipzig: 3-year fully funded postdoc intersecting climate research + text analysis + sociology
🔗https://uni-leipzig.b-ite.careers/jobposting/19cb49183f1e25db8462734e444cb5dd476862ea0

AI generated photograph featuring a group of people sat around a table and code overlayed

📣Call for papers!

1️⃣ week left to send your 250-word abstract for @audreyalejandro.bsky.social and ‪@dandekadt.bsky.social‬'s two-day workshop on "Computational Social Science meets Qualitative Research"

Learn more➡️ www.lse.ac.uk/Methodology/...

@lsedatascience.bsky.social

SODAS Lecture with Kevin Munger SODAS Lecture on the effects of TikTok on the evaluation of politicians with Kevin Munger.

Join us for the first SODAS Lecture of the spring on March 21! 📅

Kevin Munger will kick off the lecture series with a talk on the effects of TikTok edits on evaluations of politicians📱

Event🔗: sodas.ku.dk/events/sodas...