Tune into the #aBitOfCCS Podcast as we explore how large language models classify online political speech about sexism. Aditi Dutta, a doctoral researcher at the University of Exeter, joins us to discuss her study on how automated moderation systems often misclassify anti-sexist speech as harmful—raising important questions about fairness, resistance, and digital democracy.
CONTENT WARNING: This episode includes discussions and examples of sexist language online, which may be offensive or upsetting to some listeners.
Read the paper here: https://arxiv.org/abs/2508.11434v1
Reach out to Aditi at ad882@exeter.ac.uk for more insights into her research.
In this episode, we’re joined by Dr. Valerie Hasse from LMU Munich to demystify one of the most widely used tools in computational text analysis: the dictionary. Valerie explains how computational dictionaries relate (or don’t!) to the everyday dictionaries we know, and breaks down how they actually work behind the scenes. We explore what dictionaries are good for, when to build your own versus using ready-made ones, and where they shine — especially for studying opinions, emotions, and media narratives. Valerie also opens up about the real challenges that come with using dictionaries, from biases to technical hurdles, and whether they still matter in the age of large language models. She gives clear answers and practical insights into a tool that helps researchers decode massive amounts of text.
In this episode, Prof. Jamal Abdul Nasir from the University of Galway reveals why pre-processing is the backbone of all text analysis. He breaks down key steps like defining documents, tokenization, removing stop words, unification, and stemming vs. lemmatization. Jamal also explains unigrams vs. bigrams and how modern NLP techniques like byte-pair encoding are changing the game. Plus, he shares practical tips for making your pre-processing transparent and reproducible, helping your research stand strong and scale up.
In this episode, we’re joined by Dr. Fabienne Lind from the University of Vienna, who sheds light on how computational methods transform the way we study opinionated communication. Fabienne shares her experience researching political emotions on social media in the CIDAPE project and explains what it really means to “code” when we’re working with text. We explore how computational tools help us find patterns and insights that traditional reading might miss — and why this matters for understanding public discourse today. From clear benefits to real challenges, Fabienne shows why thinking computationally is key for anyone studying text at scale. Further information about the CIDAPE Horizon Europe Project here: https://cidape.eu/
Tune in to the #aBitOfCCS Podcast as we explore how to measure something as abstract and slippery as uncertainty in political speech. Ella MacLaughlin, a PhD candidate at Utrecht University, joins us to discuss her ongoing research on how politicians in the US, UK, Germany, and the Netherlands express uncertainty in public communication.
In this episode, we dive into the challenges of capturing uncertainty in political language, how it differs from other domains like biomedical science, and how we can build a dictionary for latent novel contracts.. Rather than focusing on results, we reflect on the conceptual and methodological puzzles that come with studying highly normative political language through computational tools.
Reach out to Ella at e.j.maclaughlin@uu.nl for more on her work, and check out the project website at radiunce.org!
A brief note from Ella: At 17:52, I mistakenly attributed a paper to ‘Walters’ instead of the correct author ‘Walker’. Apologies for the error.
In this episode, we are joined by Prof. Helle Sjøvaag, journalism researcher and founding member of the OPINION Network. She shares how the network came to life and why it’s vital for studying how opinions form and spread online. We explore how digital spaces — from social media to news sites — shape what we think and how we express it. Helle unpacks the hidden influence of technology, power, and money on online discourse. Tune in to hear why building supportive networks is crucial in navigating these turbulent spaces — and how collective research can make a real difference. Further information here: https://www.opinion-network.eu/about
In this episode of #aBitOfCCS, Jana Bernhard-Harrer sits down with Denise Roth, a PhD student at the Strategic Communication Group at Wageningen University & Research. Denise’s research focuses on how science is communicated by political elites and the implications for the relationship between science and society.
Her study, "In the Crossfire: Online Hostility Towards Public Figures Amid Politicized Science Communication" , investigates how large language models (LLMs) can be leveraged to annotate training data for a classifier capable of distinguishing hateful language from other types of online comments. Together, we explore the intersection of AI, social science, and combating online hostility in the context of politicized science communication.
Don’t miss this thought-provoking discussion on the challenges and opportunities in understanding and addressing hostility towards public figures in today’s digital age. If you have any questions, you can connect with Denise here: Denise Roth - Wageningen University & Research
In the first episode, Wendo King'ang'i and Christian Baden dive into the concept of opinionated communication — what it is, how it shows up in our daily lives, and why it matters. They also touch on the current state of research and what the future holds for this fascinating topic.
In this podcast we aim to introduce and discuss the OPINION COST Action CA21129. The network convenes early- and mid-career researchers from over 35 European countries, Israel, and the US, integrating cutting-edge expertise from different disciplines (notably, communication science, computational linguistics, IT) while networking the many, hitherto largely disconnected language communities of textual researchers. Further information here: https://www.opinion-network.eu/about
Join us on the #aBitOfCCS Podcast as we dive into political agenda-setting with Daniel Sandvej Eriksen, Post-Doc at the University of Aarhus, Department of Political Science.
In this episode, we discuss Daniel’s latest research, "Initiate and Elevate! How Political Parties Can Set an Agenda." His study introduces The Issue Initiation Model, which explains how political parties proactively shape discussions . . Using computational methods, Daniel analyzes over 5.5 million tweets and 750,000 news articles from the UK and Denmark (2015-2022) to uncover how parties and MPs strategically set the agenda.
Hosted by Jana Bernhard-Harrer, this episode explores the intersection of computational social science and political communication, revealing how digital traces help us understand political strategy.
In this special episode of #aBitOfCCS, Jana Bernhard-Harrer sits down with Marvin Stecker, a PhD student at the Computational Communication Science Lab at the University of Vienna. Marvin is part of the AUTHLIB project, which focuses on illiberal challenges to democracy.
Unlike our usual focus on specific computational methods, this episode explores a study Marvin conducted with colleagues from OPTED: "Tools of the Trade – When Are Software Tools Mentioned in Computational Text Analysis Research?" The study examines how software tools are reported in computational communication science research and what this means for transparency and replicability.
We discuss:
• Key findings from a review of 406 journal articles published between 2016 and 2020.
• Patterns of software tool reporting across disciplines and over time.
• How factors like methodological validation and tool accessibility influence tool mentions.
• Implications for transparency, replicability, and the future of computational text analysis in communication science.
This episode provides valuable insights into the role of software tools in shaping computational research practices and offers a critical look at how we can improve transparency in the field.
For more on Marvin’s work, connect with him at: Marvin Stecker - marvin.stecker@univie.ac.at
In this episode of #aBitOfCCS, Jana Bernhard-Harrer chats with Hung Nguyen, a research fellow at the German Institute for Adult Education – the Leibniz Centre for Lifelong Learning. Hung shares insights from his study, "A Sentiment-Based Approach to Measuring Multidimensional Party Positions with Transformer."
The study introduces ContextScale, a framework that uses AI to analyze party positions by separating political sentiments from rhetorical styles. Built on the XLM-RoBERTa model, ContextScale offers new ways to understand policy intentions and party dynamics across languages and domains. Hung also discusses the dataset released through his research, featuring party positions on 11 topics across 22 countries.
Join us as we explore the use of transformer models in political communication and the potential for reshaping how we analyze party positions at scale.
For more on Hung’s work, connect with him at: Hung Nguyen - hung.nguyen@die-bonn.de
In this episode of #aBitOfCCS, Jana Bernhard-Harrer interviews Said Unger from the University of Münster about his research on combating disinformation through prebunking interventions. Said’s study, "Immunize the Public against Disinformation Campaigns: Developing a Framework for Analyzing the Macrosocial Effects of Prebunking Interventions," explores how Agent-Based Modeling (ABM) can help identify key turning points for preventing the spread of disinformation at a societal level. Don’t miss this conversation on the challenges and innovations in combating disinformation at scale!
Additional Resources:
• Study Link: Immunize the Public against Disinformation Campaigns
• Learn ABM with Python: ABM Python Tutorial by Mariuzka
For further insights, you can reach out to Said Unger at said.unger@uni-muenster.de.
Tune in to the #aBitOfCCS Podcast as we discuss misinformation during the COVID-19 pandemic with Lotte Schrijver, a PhD candidate in Strategic Communication at Wageningen University and the Pandemic and Disaster Preparedness Center. In her study, "In Search of Truth: A Refined Approach to Collecting and Annotating Data for BERT-Based Misinformation Detection," Lotte explores the challenges of training machine learning models to detect nuanced misinformation.
This episode, hosted by Jana Bernhard-Harrer, delves into the complexities of annotation choices and the limitations of models like BERT in capturing subtle forms of misinformation, such as conspiracy theories. Reach out to Lotte at lotte.schrijver@wur.nl for further insights, and stay tuned for the DOI link to her study, coming soon!
Tune in to the #aBitOfCCS Podcast as we explore state communication and public trust during the COVID-19 pandemic with Christian Schwaderer, a PhD candidate at the University of Innsbruck, Austria. Christian’s study, Navigating Trust: Comparative Insights into Crisis Communication Strategies and Public Perception, examines how government communication strategies on social media and in press conferences influenced public perception and trust in crisis management.
In this conversation, hosted by Jana Bernhard-Harrer, we dive into Christian’s analysis of nearly 400 government press conferences across nine OECD countries, alongside real-time public reactions on social media. Using both manual coding and quantitative methods, his research reveals patterns in communication that shaped public trust and offers valuable insights for crisis response strategies.
For further discussion, connect with Christian at Christian.Schwaderer@uibk.ac.at, and you can find his slides on a presentation of this study here: https://www.slideshare.net/secret/H7CADAky5sIVwK
Credibility is a crucial concept in communication science and received severely increased attention, again, with CCS. That is, it serves everybody as a signpost to navigate the web whilst also being scrutinized by some via (AI-driven) signals that suggest trustworthiness. Cuihua (Cindy) Shen is Professor of Communication and Co-Director of the Computational Communication Research Lab at the Department of Communication at UC Davis. In this episode, she, Emese Domahidi (Professor at TU Ilmenau) and Mario Haim (Professor at LMU Munich) talk about the concept of credibility and its particular role with mis- and disinformation. Of course, we also talk AI and what credibility is worth when a machines can generate whatever we've learnt to be trustworthy.
P.S.: We now also have a website for our podcast --> https://aboutccs.net/
P.P.S.: This is the last episode of this season. We're off to a (longer? ;-)) summer pause but look forward to being in touch soon!
Ethan Zuckerman, Associate Professor of Public Policy, Communication and Information at the U of Massachusetts Amherst, is our guest, and he is on a mission to fix platforms. Not because he thinks they are inherently bad, but because there are several things about platforms that research (not least CCS) tells us are flawed. Emese Domahidi (Professor at TU Ilmenau) and Mario Haim (Professor at LMU Munich) talk with Ethan about why social media seems to be broken, what possible ways to fix it might be, how different regions of the world are approaching this challenge, and whether suing Facebook might make a difference.
P.S.: We now also have a website for our podcast --> https://aboutccs.net/
Join us in the newest episode of #aBitOfCCS Podcast featuring Dr. Mónika Simon, a Postdoctoral researcher at the UvA, unraveling Narratives of (Dis)Trust in the digital realm.
In this episode, Dr. Simon discusses her research focused on tracing information flows in contemporary media, utilizing advanced computational methods and cross-platform analysis.
Explore her paper "Linked in the dark: A network approach to understanding information flows within the Dutch Telegramsphere" co-authored with K. Welbers, A. C. Kroon, and D. Trilling.
Access the paper at https://www.researchgate.net/publication/364452085_Linked_in_the_dark_A_network_approach_to_understanding_information_flows_within_the_Dutch_Telegramsphere
For further inquiries or information, you can reach Dr. Mónika Simon atm.simon@uva.nl. Tune in to this episode for a captivating exploration of the intricate world of information flows, providing valuable insights into the digital age and the dynamics of trust and distrust in media.
Dive into the latest episode of #aBitOfCCS Podcast featuring Sean-Kelly Palicki, a PhD candidate at TU Munich, as he explores multilingual document sampling and the impact of keyword translation strategies on automated text analysis. In this engaging conversation with host Jana Bernhard, Sean discusses key findings from his study, "Selecting Relevant Documents for Multilingual Content Analysis," published in Computational Communication Research. Check out the full study at https://doi.org/10.5117/CCR2023.2.5.PALI. For further inquiries, reach out to Sean-Kelly at sean.palicki@tum.de. Don't miss this insightful episode on the nuances of document selection in computational communication research!
Everyone is talking about Artificial Intelligence (AI), so we want to bring some differentiation into the bigger picture. For this, Jean Burgess, Distinguished Professor of Digital Media in and founding director of the Digital Media Research Centre (DMRC) at Queensland University of Technology, is our guest. She has been focusing on social implications of digital media technologies, platforms, and cultures, as well as new and innovative digital methods for studying them, for quite some time and has recently become Associate Director of the national Australian Research Council Centre of Excellence for Automated Decision-Making and Society (ADMS). From that, she's perfect to discuss with us--Emese Domahidi (Professor at TU Ilmenau) and Mario Haim (Professor at LMU Munich)--about what AI really is and where the hype is coming from, what role different disciplines play and where methods come into play.
P.S.: We now also have a website for our podcast --> https://aboutccs.net/
Links
https://www.admscentre.org.au/
https://research.qut.edu.au/dmrc/
Step into the world of language model-based chatbots with our latest podcast episode! Join us for an in-depth exploration of the study titled "The Silence of the LLMs: Cross-Lingual Analysis of Political Bias and False Information Prevalence in ChatGPT, Google Bard, and Bing Chat." In this insightful episode, our host engages in a compelling interview with the researchers behind the study—Aleksandra Urman from the Department of Informatics at the University of Zurich (urman@ifi.uzh.ch) and Mykola Makhortykh from the Institute of Communication and Media Studies at the University of Bern (mykola.makhortykh@unibe.ch).
Discover key findings from their groundbreaking research, offering a cross-lingual analysis of political bias and false information prevalence in large language model-based chatbots. Uncover the implications of their work on the trustworthiness of AI-driven chat systems.
For further inquiries or to join the conversation, reach out to Aleksandra and Mykola via email. This episode provides a thought-provoking journey into the complexities of language models, political bias, and the prevalence of false information in the realm of contemporary chatbot technologies. Access the full study here: https://osf.io/q9v8f/download
