Seed Funding
Through our Seed Funding projects, we facilitate research focused on the application and innovation of digital methods. Find more information on the projects funded by the lab below:
2024
By Saurabh Khan, Irene van Driel, Sindy Sumter, Chei Billedo, Lauren Taylor, and Olga Eisele
Our research introduces PictoPercept, an innovative open-scientific, web-based visual survey method designed to assess societal biases and perceptions in a non-obtrusive manner, with our first implementation focusing on youth perceptions of occupational roles. By leveraging the concept of pairwise wiki surveys and integrating visual stimuli, PictoPercept offers a nuanced tool for capturing sensitive societal perceptions without the influence of social desirability bias that often plagues traditional survey methods. This method allows participants to both respond to and contribute survey questions, facilitating a collaborative, adaptive, and inclusive data collection process. Initial applications target Dutch youth’s occupational stereotypes, and can reveal PictoPercept’s potential to uncover the intricate dynamics of digital media’s impact on societal perceptions and occupational stereotypes, promising significant contributions to communication science and beyond.
2024
By Priska Breves, Marthe Möller, and Jan-Philipp Stein
Social media platforms confront individuals with diverse body-focused content on a daily base, triggering social comparison processes that profoundly impact well-being, body satisfaction, and health-related behaviors. However, traditional self-assessment scales often fall short in capturing these automatic processes. This project aims to investigate the feasibility of using eye tracking as a novel method to measure social comparison intensity. By conducting a lab experiment manipulating influencer body shapes and user comments on Instagram posts, we discern when and where individuals direct their attention within a social environment and assess the predictive ability of eye tracking compared to self-report measures.
2024
By Justin Chun-ting Ho & Chung-hong Chan
The study examines the challenges and strategies for analyzing nationalist frames in multilingual and multinational texts, with a special focus on linguistic and contextual transferability. Measuring nationalism from text is a challenging task. While pioneers have made efforts to capture nationalist frames and discourse using computational methods, these attempts are limited to single country and single language applications. Due to the nuanced and context sensitive nature of nationalism, approaches that work in one case are not guaranteed to work in another social setting. This project aims to address these limitations by developing a comprehensive framework for identifying nationalist frames across diverse languages and social contexts.
The project uses political advertisements from 90 countries, manual translation, and crowd-annotation. The performance of three large language models (LLMs)—Multilingual BERT, LLaMA, and Mixtral—will be evaluated for their ability to adapt across languages and contexts. Findings will enhance computational frame identification and contribute to understanding nationalism globally.
2024
By Felicia Loecherbach & Laura Boeschoten
Messaging services such as WhatsApp potentially play a role for sharing political information via meso-spaces between public and private. However, it is difficult to tell in how far WhatsApp IS being used for news, especially in European contexts since data access so far has been very tedious and poses many privacy-related challenges. Data donation allows researchers to access such otherwise inaccessible data sources. In this project, we combine two existing open-source infrastructures: WhatsApp Explorer to help respondents retrieve their own data and PORT to donate that data to researchers. In a pilot study, we investigate whether and if so which links to news are being shared via WhatsApp.
2024
By Lina Buttgereit, Michael Hameleers, Damian Trilling, Katjana Gattermann, and Andreas Schuck
Despite analyses suggesting that mis- and disinformation make up only a small share of citizens’ daily news diets, the perceived misinformation exposure and worries from audiences remains high. We explore to what extent news users employ a more encompassing definition of misinformation compared to the academic literature, following recent evidence suggesting that misinformation discourse, individual biases, and disagreement on the identification of misinformation could affect misinformation perceptions. Using an ESM study (N = 1000), we employ an audience-centered approach to map organic perceived misinformation encounters and motivations for perceived misinformation encounters. In doing so, we explore the alignment between perceived misinformation exposure and existing academic approaches.
2023
By Frederic Hopp and Linda Bos
This project investigates the use of moral foundations by Dutch political elites on social media. We will collect all social media posts from Dutch party leaders and parties posted between January 2021 and May 2023, and subsequently use crowd-coding to obtain annotated moral foundations on a subset of these posts. We then fine-tune a cross-language BERT model (XLM-R) on a corpus of English Tweets annotated for moral foundations and test how well this model can classify moral foundations in our crowd-sourced Dutch social media posts. In addition, we aim to computationally explore and classify the visual cues that accompany and undergird the moral language of Dutch political elites.
2023
By Gian Hernandez
This study explores diverse embodiment through the use of innovative digital methods, focusing on podcasting as a medium for science communication (O’Hara, 2020). Criticality has emerged as a vital issue in health and fitness communication; notions of structure and agency remain undertheorized in the field (Lupton, 2009; Tiller et al., 2022; Zoller et al., 2019). The project utilizes a website for both dissemination and research purposes, on which podcast interviews regarding embodiment, diversity, and health and fitness with prominent experts from diverse backgrounds will be hosted. Transcripts of the recorded conversations will be posted on the platform to be commented on annotatively by audiences interested in critical health and fitness. Finally, these comments on the transcribed conversations will be analyzed using critical discourse analysis (Wodak & Meyer, 2001) to uncover audience attitudes, perceptions, and engagement with diverse themes within critical health and fitness.
2023
By Zilin Lin, Susan Vermeer, and Anne Kroon
Machine learning has been thriving in the field of communication science. Yet, it should be noted that decent model performance could only be possible to achieve when there is an ample amount of training data with correct annotation. Such model input, unfortunately, is sometimes difficult to obtain, due to the trade-off between quality and quantity within a reasonable research budget and timeframe. In our study, we would like to explore the potential of crowdsourcing as an approach to providing accurate model input. Specifically, we investigate whether annotation biases exist, and if so, whether they are associated with different individual characteristics.
2023
By Marthe Möller and Joanna Strycharz
Comments written in response to social media content can tell a lot about how people experience this content. The present project uses social media comments to detect users’ entertainment experiences in response to social media messages. It does so by analyzing the comments posted in response to emotional corporate films in particular. This way, the project adds to the methods that scholars have to measure viewers’ experiences of online entertainment content. In addition, by comparing this novel way of measuring entertainment experiences to more established methods for measuring entertainment experiences (i.e., surveys), the project advances our methodological understanding of different approaches to studying entertainment experiences.
2023
By Carolin Ischen, Theo Araujo, Jochen Peter, and Alain Starke
Conversational agents (CAs) can make personalized product- or service-related recommendations based on user input, and allow for repeated interactions with their users over time. This study distinguishes between within-session effects which refer to the (longitudinal) effects of one-shot personalized recommendations, and between-session effects which refer to the effects of a CA remembering user input from previous interactions (conversational memory). We test the persuasive effects of these two types of personalization. This project makes a methodological contribution: It extends our conversational agent research toolkit by (1) integrating recommender systems and (2) working with conversational memory over time.
2022
By Sarah Marschlich
Previous research conceived social media affordances as opportunities for action enabled by platforms technologies or as the result of users’ agency, neglecting that users act building on their perception of platform properties, gratifications, and other individual factors. To overcome the shortcomings, this research conceives affordances as the interplay between the social media platforms and the individual user, analyzing affordances in the context of organizations’ communication on social media. Combining survey, data donation, and in-depth interviews in which individuals are exposed to their actual social media usage, this project sheds light on the emergence of social media affordances and their consequences for user engagement, considering technological features and individual factors. The methodology allows for a much more valid and accurate measurement of social media affordances and significantly contributes to collecting digital communication data through donated log-based data and self-report measures.
2022
By Marijn Meijers, Jeroen Lemmens, and Hanneke Hendriks
Climate change is posing an increasing threat to nature, animal species, and humans. To combat climate change, individuals play an important role by behaving climate friendly and by engaging in interpersonal communication regarding climate change in order to create support for societal and governmental changes. Research so far mostly focuses on stimulating climate friendly behavior, whereas stimulating interpersonal climate change communication is so far largely overlooked. The question, therefore, is: How can we stimulate people to talk about climate change? We investigate whether a Virtual Reality experience regarding climate change stimulates interpersonal climate change communication. Furthermore, we look into the role of emotions and arousal as underlying mechanisms.
2022
By Sonia Jawaid Shaikh
Artificial intelligence (AI)-enabled technologies are increasingly being used by humans to make decisions across a variety of settings. The basic idea underlying this technological setup is to provide humans with some kind of machine-based recommendation (e.g. employability score, risk score) which can be utilized in making a judgment or decision. An important consideration in these circumstances has to do with overconfidence i.e., the extent to which humans systematically overestimate accuracy of their decisions and precision of their knowledge. Overconfidence negatively affects various organizational outcomes. However, it remains unclear how and to what extent it evolves when human decision-making is influenced by machine recommendations. This project investigates antecedents, extent, and correction of overconfidence within the context of machine-driven recommendations for decision-making. A series of experiments on a virtual platform will investigate how interaction with AI affects overconfidence and also explore design-based solutions to reduce it.
2022
By Priska Breves, Sophie Boerman and Nicole Liebers
Social media influencers who post body-focused content have often been criticized for their negative impact on young women (e.g., regarding eating disorders). However, empirical research is lacking, especially concerning the long-term effects. By following and engaging with the same influencer over several weeks, followers are likely to form strong one-sided illusionary relationships. These parasocial relationships have been connected to an intensification of social comparison processes. Thus, the influencer’s impact on participants’ body perception, well-being, and health-related behavior should also be enhanced after several weeks. In our study, we use Experience Sampling and confront participants with daily posts from the same influencer while experimentally analyzing the impact of the influencer’s thematic focus and body type.
2022
By Lara Wolfers and Susanne Baumgartner
With the deeper integration of digital media into everyday life, the popularity of Experience Sampling Methods (ESMs) in communication science has surged. For ESMs, it is essential to keep the length of daily questionnaires as short as possible. Thus, constructs are often assessed with one item. However, even for key communication constructs, no validated one-item measures are available. The aim of this project is to validate one-item measures for key communication constructs for use in ESM. We will identify central communication science construct, select suitable items for these constructs, and validate these items in an ESM study.
2022
By Marthe Möller, Susan Vermeer, and Susanne Baumgartner
The past years have seen a rise in studies investigating how user comments influence the experiences of social media users. The various studies that have been conducted on this topic so far all seem to be based on the same assumption, namely that comments are relevant in the sense that they actually discuss the main content that they accompany (e.g., a YouTube video, Instagram image). The present project tests this assumption by creating a tool that can detect the relevance of comments (i.e., whether or not a comment discusses the main content that it accompanies). Using comments written in response to music videos posted on YouTube, we train a machine that can automatically classify comments. Scholars can use this machine to verify the assumption that the comments that they study are relevant (and hence, can influence how viewers experience social media content).
This project is reported in the following publication: Möller, A. M., Vermeer, S. A. M., & Baumgartner, S. E. (2024). Cutting through the comment chaos: A Supervised Machine Learning approach to identifying relevant YouTube comments. Social Science Computer Review 42(1), 162-185. https://doi.org/10.1177/08944393231173895
2022
By Irina Lock
Anecdotal evidence shows that news articles visualize AI repeatedly as stylized humanoid robots or brains, speaking to technology’s anthropomorphization. While AI frames have been studied in news texts, content analyses of AI images are lacking. The way news media visually frame AI purports multiple sociotechnical imaginaries of how the reader is supposed to envision the future. However, we miss systematic knowledge about how news media visually frames AI. This project uses images from open source and commercial editorial image databases to build a classifier that categorizes pre-defined frames in images of AI to be applied to analyses of news or social media.
2022
By Christel van Eck, Anne Kroon, and Damian Trilling
While increasingly more studies focus on online climate change polarization dynamics, research on ‘processes’ of polarization (instead of ‘states’ of polarization) and research on depolarization dynamics is largely overlooked. Hence, the current study investigates online climate change polarization and depolarization dynamics, by analyzing Dutch comment interactions about climate change on Youtube and Twitter. A sample of comments is manually annotated by crowd coders based on the following concepts: (a) users’ global warming stance; (b) escalation; and (c) identity labeling. Using these annotations, we train a BERT supervised classifier to predict class membership in a large-scale dataset of social media comments on the topic of climate change. This project provides insight into how climate change polarization emerges, accelerates, and dissolves online.
2022
By Joanna Strycharz and Joseph Yun
Studying attitudes has been crucial in communication and advertising theory development; measurement of attitudes has had a substantial impact on several contexts, including research on advertising effectiveness and consumer reactanceIn recent research, it has been acknowledged that while attitudes are important for beliefs and behavior, strong attitudes have a greater impact on individual’s intentions and behaviors. Thus, both attitude and attitude strength take an important role in communication and advertising research. In this study, we investigate how attitude and attitude strength measurements can be improved in the context of digital communication.
2021
By Anne Kroon and Toni van der Meer
Algorithms are fundamentally transforming how organizations recruit job candidates. The current project investigates the extent to which algorithmically-driven resume search engines inhibit or facilitate gender and age inequality in the recruitment process. The novelty of the project lies in tracing the influence of hidden (in addition to manifest) features that may implicitly signal social membership, such as variations in writing style, work experience, and hobbies listed. The potential for inequality is especially critical for hidden features—as they are arguably more difficult to identify and may therefore affect ranking despite explicit efforts to debias training data and algorithms.
2021
By Brahim Zarouali and Tom Dobber
This project adopts psychological profiling by predicting consumers’ personality based on their text from Twitter. This is done with a machine learning (ML) algorithm. In addition, the project investigates the validity of this personality classifier. This is achieved by comparing the algorithmic scores to self-reported personality scores. Taking the self-report as a golden standard, we are able to draw conclusions regarding the accuracy and validity of the algorithm in predicting people’s personality based on social media data. As such, it can prove to be a valid substitute for personality assessment.
2020
By Hande Sungur & Jessica Piotrowski
While public awareness for environmental protection has been rising, individuals still resist making necessary changes in their daily lives to help. Promoting sustainable behaviour is challenging because consequences of not engaging in these behaviours (e.g., climate change, plastic pollution), tend to be psychologically distant from our reality. Virtual Reality (VR) technology provides the opportunity to bring psychologically distant events closer and allow audiences to engage in important but typically “distant” topics such as sustainability in a unique way. While there is great interest in integration of digital tools such as VR in education, especially in the field of sustainability, there is still much that we do not know about using immersive environments with children. This project by taking into developmental considerations into account aims to improve our understanding of designing effective immersive experiences for children. Specifically, we design a developmentally appropriate VR environment that aims to educate children about plastic pollution in oceans.
2020
By Marijn Meijers, Anke Wonneberger, Heather Torfadottir, and Ewa Maslowska
Climate change is one of the biggest challenges of our time. Once people experience the consequences of climate change (e.g., a forest fire or a flood), they are more likely to perceive climate change as an actual threat and are, therefore, more likely to act in a pro-environmental manner. We posit that Virtual Reality (VR) offers an innovative manner to allow people to experience the consequences of climate change (e.g., experiencing a forest fire). In this project, we investigate how both cognitive and emotional responses can be triggered by such a VR experience, and how this subsequently affects behavioral intentions and actual behavior.
This project is reported in the following publication: Meijers, M. H., Torfadóttir, R. H., Wonneberger, A., & Maslowska, E. (2023). Experiencing climate change virtually: The effects of virtual reality on climate change related cognitions, emotions, and behavior. Environmental Communication, 17(6), 581-601. https://doi.org/10.1080/17524032.2023.2229043
2020
By Mark Boukes, Anne C. Kroon, and Theo Araujo
An often-researched question is whether online debate between citizens lives up to the democratic standards of a deliberative public sphere (e.g., Freelon, 2015; Jaidka, Zhou, & Lelkes, 2019). However, previous studies remained rather descriptive on the presence of certain speech elements. We instead want to investigate the following research question: How do online discussions take shape and what kind of user-comments may evoke subsequent comments of a higher/lower deliberative quality? To answer this RQ, we analyze the content of these posts using supervised machine learning and statistically model how discussions evolve over time to predict how certain types of response may influence the nature of later posted user-comments.
This project is reported in the following publications:
- Boukes, M., Chu, X., Noon, M. F. A., Liu, R., Araujo, T., & Kroon, A. C. (2021). Comparing user-content interactivity and audience diversity across news and satire: differences in online engagement between satire, regular news and partisan news. Journal of Information Technology & Politics, 19(1), 98–117. https://doi.org/10.1080/19331681.2021.1927928
- Boukes, M. (2024). Deliberation in online political talk: exploring interactivity, diversity, rationality, and incivility in the public spheres surrounding news vs. satire. Journal of Communication, jqae038. https://doi.org/10.1093/joc/jqae038.