Survey research is often interpreted as showing that belief in conspiracy theories can be surprisingly widespread, including belief in conspiracy theories that would be astonishing if true. For example, in The Atlantic we learn that “12 million Americans believe lizard people run our country”
Someone compiled all of the Calvin and Hobbes strips about Calvin’s polls of his Dad, which is proof that good people do exist on the internet
imgur.com/a/poll-resul...
![2026-04-22 | Input talk | Leah von der Heyde (GESIS)
AIn't Nothing But a Survey? Using Large Language Models for Coding Open-Ended Survey Responses
Hybrid event [A5, 6, Room A231 + Zoom]
April 22, 2026, 13:45-15:15
Abstract
Large language models (LLMs) have the potential to make survey research more efficient, including the classification of open-ended survey responses. However, as most existing research on this topic has focused on English-language text or single LLMs, it is unclear whether their applicability generalizes and how the quality of classifications compares to established methods. In this talk, I will demonstrate how LLMs can be used for coding open-ended responses using different access options and prompting and fine-tuning techniques. I will present a study testing these approaches on a dataset of German open-ended survey responses, comparing several LLMs to human coders and other automated methods. Finally, I will discuss the implications of the study findings for practitioners, including the many trade-offs researchers need to consider.
Presenter(s)
Leah von der Heyde is a computational social scientist and survey methodologist. Her research focuses on the potential and pitfalls of new data sources, such as large language models, for improving the measurement and representation of public opinion. Substantively, she is particularly interested in political attitudes and voting behavior. Leah received her PhD in Social Data Science and Research Methodology from the University of Mannheim. She has a background in political science from LMU Munich, the University of Mannheim, and Georgetown University. Previously, Leah was part of the Social Data Science and AI Lab at LMU Munich, worked for the European Social Survey at GESIS - Leibniz Institute for the Social Sciences, the European Parliamentary Research Service, and several market and public opinion research institutes in Germany and Sweden. At GESIS, she is part of KODAQS, researching and educating…](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:oizrkjzyr6vekl76tjctghcl/bafkreigb4xgjzges3yeoup7w3z3goxg6fgei4qiystwxbq4zkvjughks6m)
2026-04-22 | Input talk | Leah von der Heyde (GESIS) AIn't Nothing But a Survey? Using Large Language Models for Coding Open-Ended Survey Responses Hybrid event [A5, 6, Room A231 + Zoom] April 22, 2026, 13:45-15:15 Abstract Large language models (LLMs) have the potential to make survey research more efficient, including the classification of open-ended survey responses. However, as most existing research on this topic has focused on English-language text or single LLMs, it is unclear whether their applicability generalizes and how the quality of classifications compares to established methods. In this talk, I will demonstrate how LLMs can be used for coding open-ended responses using different access options and prompting and fine-tuning techniques. I will present a study testing these approaches on a dataset of German open-ended survey responses, comparing several LLMs to human coders and other automated methods. Finally, I will discuss the implications of the study findings for practitioners, including the many trade-offs researchers need to consider. Presenter(s) Leah von der Heyde is a computational social scientist and survey methodologist. Her research focuses on the potential and pitfalls of new data sources, such as large language models, for improving the measurement and representation of public opinion. Substantively, she is particularly interested in political attitudes and voting behavior. Leah received her PhD in Social Data Science and Research Methodology from the University of Mannheim. She has a background in political science from LMU Munich, the University of Mannheim, and Georgetown University. Previously, Leah was part of the Social Data Science and AI Lab at LMU Munich, worked for the European Social Survey at GESIS - Leibniz Institute for the Social Sciences, the European Parliamentary Research Service, and several market and public opinion research institutes in Germany and Sweden. At GESIS, she is part of KODAQS, researching and educating…
🚨 Upcoming: "AIn't Nothing But a Survey? Using Large Language Models for Coding Open-Ended Survey Responses"
👤 Leah von der Heyde (GESIS)
🗓️ Wed, April 22, 13:45-15:15 CET
📺 Register for the live stream: us02web.zoom.us/meeting/regi...
🔗 socialsciencedatalab.mzes.uni-mannheim.de/page/events/
Also finding it hard to keep up with new research? I built something to fix this.
SciLove — swipe through recent papers in your field. The feed learns from your saves. Also matches you with researchers saving your work back (opt-out if you prefer).
www.scilove.app
3,000+ journals, updated daily
⏰ Last chance to register for my GESIS course
▶️ Bayesian Modeling: From Foundations to Custom Solutions
📅 Apr 16–17 & 23–24 (online)
We cover:
• Foundations of Bayesian inference
• Applied modeling with brms in R
• Building custom models in Stan
Feel free to get in touch if you have questions.
🧵on my new paper "Synthetic personas distort the structure of human belief systems" w Roberto Cerina I'm v excited about...
🚨 Do synthetic samples look like human samples?
We compare 28 LLMs to the 2024 General Social Survey (GSS) to find out + develop host of diagnostics...
Join us tomorrow for this exciting talk by @klaramueller.bsky.social in the MZES Social Science Data Lab!
Details and Zoom link ⬇️
We are delighted to welcome Leah von der Heyde! Leah joined KODAQS as a postdoc in the SDM department at GESIS, focusing on data quality aspects of survey data using AI methods. Leah's research explores AI usability for measuring and representing public opinion.
#KodaqsAcademy #AIinSurveys
We‘re very excited for the next edition of the Summer School for Women* in Political Methodology in Mannheim! ✨
Join us and apply until 1 March!
More info in @melinscribe.bsky.social‘s thread below and here: summerschoolwpm.org
Application form: sosci.sowi.uni-mannheim.de/wpm/
🚨 #Bots and #LLMs threaten the integrity of online surveys and public opinion research, but we can identify LLM-generated text in open narrative responses by fine-tuning #BERT.
New #OpenAccess article with @jkhoehne.bsky.social @rubac.bsky.social @carohaensch.bsky.social.
👉 doi.org/10.1177/0894...
Personal update: proud owner of the first pair of MZES merchandise socks. Being a staff representative @mzesunimannheim.bsky.social for four years was totally worth it! 💪😅
this shit is just beyond parody, man
That’s an awesome project - thank you!
My Quarto #rstats tip for big/complex papers is to set up R scripts that do data processing/modeling & save plots/outputs as RDS files. The Quarto can then load the RDS file & make any modifications to the ggplot2 object before plotting. This avoids unwieldy long Quarto docs /w complex caching.
9.2% of all citations go to just 0.32% of psychology articles.
To have the most impact, post-publication peer-review should focus on these influential articles with ≥30 citations per year.
@error.reviews will systemically sample from such articles.
Blog post:
mmmdata.io/posts/2025/0...
Autofreie Quadrate & Bäume auf alle Parkplätze 💁♀️
Habt ihr schon vom größten Tik Tok-Datensammelprojekt in D gehört?
@weizenbauminstitut.bsky.social, @dizh.bsky.social, @br24.de & @stuttgarter-nachrichten.de haben For You-Page von ca. 700 Nutzer:innen ausgewertet. 📲
Spannend: Filterblasen-Theorie auch hier wiederlegt
www.br.de/nachrichten/...
Attending this amazing workshop on data donations by @valeriehase.bsky.social @frodewald.bsky.social 👩🏻💻
#COMPTEXT2025
A stimulating morning of hands-on tutorials at #COMPTEXT2025 introduced participants to advanced techniques in digital trace data collection, visual content analysis, and Bayesian text modelling. The day is far from over — stay tuned for more cutting-edge methods in action.
holy hell this is smart
This is very interesting and more data-driven than you might think from the title. Also comes up with an actual idea what centre-left parties could do to build a new coalition. No idea if this would work, but at least it's something constructive #farright #Germany #btw25
Nice :D - There’s always next year to submit this :)
Kick-off for our DFG project! With multimodal ML, we study 15,000h of German parliamentary speeches. Stay tuned for how MPs handle populism, how facial expressions send coalition signals and dynamics of salience. #DreamTeam #ScienceRocks #DankeDFG @pluggedchris.bsky.social @ankuepfer.bsky.social
Our systems now automatically track online ad spending on Meta and Google by 109,592 advertisers from 729 parties in 51 countries. More at trends.whotargets.me
Cool !:) Using R on Fedora too with a system copr and it works way smoother and faster - can definitely recommend it. There seems to be something similar for apt too: thelinuxcode.com/install-r-an...
By using install.packages() you will always have to build each package yourself which takes time…
A screenshot of the call for applications.
🚨 CfA: MZES-DVPW Conference "Methods of Political Science"
🗓️ Mannheim, Mar 27-28, 2025
⏰ Deadline: Dec 16, 2024
🔗Apply via sosci.sowi.uni-mannheim.de/mzes-dvpw-me...
@gessler.bsky.social, @aleininger.bsky.social, @hannahrajski.bsky.social, Oliver Rittmann & I look forward to your submissions!
The G.I. Bill, Standardized Testing, and Socioeconomic Origins of the U.S. Educational Elite Over a Century Ran Abramitzky, Jennifer K. Kowalski, Santiago Pérez, and Joseph Price NBER Working Paper No. 33164 November 2024 JEL No. 123, 124, N32 ABSTRACT We compile, transcribe, and standardize historical records for 2.5 million students at 65 elite (private and public) U.S. colleges. By combining these data with more recent survey and administrative data, we assemble the largest dataset on the socioeconomic backgrounds of students at American colleges spanning the last 100 years. We document the following: First, despite a large increase in the share of lower-income students in the overall college-going population, the representation of these students at elite private or public colleges has remained at similarly low levels throughout the last century. Second, the representation of upper-income students at elite colleges decreased after World War II, but this group has regained its high representation since the 1980s. Third, while there has been no increase in the economic diversity of elite private and public colleges, these colleges have become more racially and geographically diverse. Fourth, two major policy changes in the history of American higher education, namely the G.I. Bill after World War II and the introduction of standardized tests for admissions, had little success in increasing the representation of lower- and middle-income students at elite colleges.
Holy crap this is an astounding piece of historical research
Will post ungated link later today unless someone beats me to it
www.nber.org/system/files...
Two maps the the contiguous United States. The first is colored with the 4 time zones. Above that map states "say no to time zones". The bottom is the same map but with a continuous color gradient. Above that map states "say yes to the time gradient"
We must stand against the arbitrary categorization of continuous variables!
... and that's why I'm proud to announce my support of abolishing time zones in favor of the time gradient
Gestern abend war ich ja noch sehr spät auf der Hotelterasse, um etwas den Südhimmel zu fotografieren. Bei Objekten, die sehr hoch am Himmel stehen, stößt das Seestar S50 leider an seine Grenze, und bekommt die Nachführung nicht mehr so gut hin. Während ich darauf gewartet habe, dass meine 1/
Ich finde ja, mit solchen Qualifikationstellen hat sich die Uni als Qualifikationsort disqualifiziert. #IchBinHanna
