Method Preconference Stream 1

Room U42, Faculty of Social Studies, Brno (Czech Republic)

9:30–11:00

Workshop 1: Analyzing Social Media Texts through Co-occurrence Analysis
This workshop provides an introduction to conducting co-occurrence analysis using the programming language R. Automated approaches to text analysis are essential for managing large corpora and o=er crucial insights that can complement or precede manual interpretation (Günther & Quandt, 2016). Co-occurrence analysis can be broadly understood as a largely unsupervised, automated method for measuring distance, relationships, and dependencies between words that appear within the same text document, corpus, or defined window (Momtazi et al., 2010).

As an inductive approach, co-occurrence analysis enables researchers to extract patterns from the material itself. It identifies how words (or hashtags) appear together within predefined text segments and models these relationships systematically. In this workshop, participants will learn how to conduct co-occurrence analysis step by step: from cleaning and preparing textual data, to creating co-occurrence matrices, to visualizing results as networks, in which words function as nodes and their co-use is represented as weighted edges, and finally to interpreting these structures.

From a data-driven perspective, co-occurrences provide empirical predictions of a word’s “neighborhood.” Linguistically, they often reflect syntactic or grammatical relations, such as idiomatic expressions (Evert, 2009). In communication studies, co-occurrences are typically interpreted as associative or interpretive frames (van Atteveldt, 2008). Adopting a perspective rooted in critical communication studies, this workshop treats social media text as a network of meaning. This approach enables researchers to not only analyze large-scale processes of meaning-making, but also, identify attitudes, emotions, and sentiments toward political issues, and interrogate questions of power relations, antagonism, and alignment (Makhashvili, 2024).

Presenter and Bio:
Ana Makhashvili (she/her) is a postdoctoral researcher in media and communication studies. She currently works at the Collaborative Research Center Affective Societies at Freie Universität Berlin on the research project “Contested Order of Emotions: (Anti-)Feminist Discourses on Social Media.” Her research focuses on the role of affect and emotions in far-right and anti-feminist mobilizations on social media. Her methodological expertise lies in mixed-methods research designs, combining network analysis, automated text analysis, and critical discourse analysis.

11:00–11:15
Coffee break (provided)

11:15–13:00
Workshop 2: Using machine learning/AI methods for text classification in communication research
Automated text classification has been a rapidly growing trend in the communication sciences, allowing researchers to scale up content analysis research designs to increasingly large samples. This trend has accelerated with the popularity and accessibility of generative language model such as ChatGPT, meaning that automated text classification has lower barriers in relation to computational methods knowledge or requirements labelled training datasets (zero-shot). However, use of these models also raises new questions regarding validity, reliability, reproducibility, and digital autonomy in content analysis research. This workshop will bring together communication researchers interested in text classification to share best practices and use-cases, as well as discuss the challenges of machine learning/AI classification for content analysis.

The workshop has two parts. Part 1 (~75 minutes) focuses on practice: participants briefly present a current or planned classification task (research question, unit of analysis, categories), then we compare methodological choices (supervised vs. zero/few-shot; dictionaries vs. embeddings/LLMs) and discuss best practices for documentation, prompt/model selection, reliability (e.g. output validation vs. intercoder reliability), and bias/robustness checks. In the second part (~30 minutes), João Gonçalves will share the findings of his NWO Veni research project on data quality and language models for text classification in communication research, which, together with the outputs of the first part, will result in a curated resource document and contact list for workshop participants.

Presenter: João Gonçalves, PhD, Associate Professor AI and Digitalization, Rotterdam / The Netherlands

13:00–14:00
Lunch
While lunch isn't provided, you'll find a great variety of dining spots just a short walk away. Whether you're looking for a quick bite or a sit-down meal, there are plenty of choices nearby—including excellent vegan options.
Explore these local favourites within easy walking distance:

Naše Vaše Bistro (4 min walk)
https://nasevasebistro.cz/

Vegan AF Ramen Brno (3 min walk)
https://veganaframen.cz/

Desi Dhaba (Indian, 2 min walk)
https://www.instagram.com/desidhaba.cz/

Blue Demon Bistro (Mexican, 3–4 min walk)
https://www.facebook.com/bluedemonbistro

Zdravý Život (4 min walk)
https://www.zzbrno.cz/

14:00–16:00
Workshop 3: Simulating Chaotic Information Environments
At this simulation workshop we will explore how to use simulations for data collection in various academic disciplines. Serious games can be used to mimic complex, sometimes chaotic real-life scenarios that are otherwise difficult for researchers to measure or reach. We will also shortly discuss and exemplify how to simulate qualitative metrics (aka pseudoquantification, and how to replace sensitive (e.g. personal) or difficult-to-access data with synthetic data. In addition, we will learn the use of large language models in the design of dynamic and challenging simulation scenarios.

In two hours, we will first dedicate approximately 45 minutes to go over the theoretical basis of simulating communication – a phenomenon brutally simple, yet incredibly complex to measure without total control of the environment. However, control of the environment is not necessarily essential if we know exactly what to measure, and what that measure tells us. For this workshop simulation, our central measure will be the perceived situation awareness. After the introduction we’ll divide the participants into role-based groups and play through a dynamic crisis scenario, parallelly showcasing the tools needed for collecting research-grade data and facilitating a communication simulation. The remainder quarter hour is for wrap-up, cooldown and reflections.

Presenter: Dr. Sten Torpan, Lecturer in Crisis Sociology, University of Tartu, Estonia. sten.torpan@ut.ee
https://www.researchgate.net/profile/Sten-Torpan

Bio: His work bridges classical sociology and communication research with computational social science in the domain of disaster research and communication studies. He’s focused on developing training methods and scientific simulation methodologies to improve risk and crisis communication practice.

16:00–16:15
Coffee break (provided)

16:15–17:45
Workshop 4: Bridging participatory filmmaking and film studies: Applications, challenges and future paths
This workshop explores how participatory filmmaking (PF) can be strengthened through analytical and conceptual tools drawn from film studies by highlighting its cinematic capacities as a mode of knowledge production. Participatory filmmaking, or participatory video (PV), is a set of methods that involve a group or community in shaping and creating their own film (Lunch and Chris Lunch, 2006). It has a long genealogy in education, health, development studies, agricultural extension, and applied research. While this history has enabled powerful forms of speaking “with” instead of “about” communities (Lenette, 2019), it has also contributed to PF being treated primarily as an instrument for policy-oriented communication. As a result, both the filmmaking product (films) and process (workshops) are often under-theorised, the former as narrative and aesthetic form, and the latter as a sequence of cinematic decisions.

This workshop responds to this methodological gap by bringing PF into dialogue with film studies and understanding participatory films as an embodied and relational method than engage with blocks of time and space through which knowledge is organised, sensed, and communicated. It offers a reflexive, methods-oriented toolkit for researchers who use (or plan to use) PF within qualitative inquiry. It addresses key challenges raised in critical scholarship such as the gap between participatory ideals and practice; the balance between collaborative and identity work; authorship, ownership, and consent; visibility versus exposure; extractive dynamics within asymmetric structures; and the tendency to privilege process over output. Rather than abandoning PF’s ethical commitments, the workshop proposes a combined stance: (1) a processual perspective attentive to collaboration and power relations; and (2) a textual/aesthetic perspective attentive to narrative, framing, sound, performance, and montage. The session is primarily conceptual and discussion-based, supported by short examples and a brief, optional smartphone exercise. It concludes with practical guidance for strengthening PF research designs that expands its cinematic and analytical potential not only as a representational tool, but also as a powerful epistemological tool.

Presenter: Dr. Irene Gutiérrez-Torres
Echo, Vrije Universiteit Brussel and Tecmerin, University Carlos III of Madrid
Irene.Gutierrez.Torres@vub.be
https://orcid.org/0000-0003-0563-9295