Digicom Thesis Grant Awardee Alicia Gilbert wins Top Paper Award

Former research master student Alicia Gilbert has been awarded a top-paper award at the conference of the Media Reception and Effects division of the German Association of Communication (January, 2021). Alicia received this award for a paper based on her master thesis project that was funded by the digicom lab “Online vigilance and goal conflict stress smartphone users out: An in-situ approach to digital stress”. The thesis was supervised by our digicom lab researcher Dr. Susanne Baumgartner. Alicia is currently a PhD candidate at the Communication Science department at the Johannes Gutenberg University in Mainz, Germany. We congratulate Alicia for this great achievement!

The call for the fifth edition of the Digicom Thesis Grant Awards is now online.

New Master Theses funded by digicomlab Grant

Four new Master Theses receive digicomlab Funding Grant

The Research Priority Area Communication and its Digital Communication Methods Lab are happy to announce that four proposals have been selected for the fourth edition of the digicomlab Thesis Funding Grants. These grants provide financial support for theoretically-relevant and digitally innovative (research) master theses written at the University of Amsterdam’s Graduate School of Communication for semester 1 of the 2020-2021 academic year.

Deepfake Negativity: Political attacks, disinformation and the moderating role of personality traits on evaluations of politicians

Monika Simon @si_moni_ka

Deepfakes are artificial intelligence (AI) enabled doctored videos that strongly resemble real videos. While scholars worry that deepfakes represent a new powerful form of disinformation (Dobber et al., 2020), which may enable malicious actors to poison the public debate and interfere with democratic elections (Vaccari & Chadwick, 2020), only two studies to date have explored the credibility and attitudinal impact of political deepfakes. To fill this gap in literature, the present study employed a self-produced deepfake of a politician which was created by a pre-trained deep learning model that utilized 2D expression swapping to modify the lip shape of the politician frame by frame.

Synthesizing literature on online disinformation, negativity bias, and negative campaigning, the study will uncover how are voters’ political attitudes affected by exposure to a deepfake featuring attack politics, which has become a prominent characteristic of contemporary US politics. Using an online experiment among US citizens, the credibility and attitudinal impact of a self-produced deepfake featuring an uncivil character attack sponsored by a Democrat targeting a Republican will be compared to real footage featuring a policy attack of the same sponsor and target. As initial evidence shows that effectiveness of negative campaigning “may be a matter of taste” (Nai & Maier, 2020), the present study uncovers the interaction between negativity and personality traits which may contribute to clarifying the so far inconsistent effects of negative campaigning. In sum, the present study will advance knowledge on the attitudinal impact of negative campaigning and disinformation by employing an AI-powered self-produced political deepfake.

 

Acquiring Political Knowledge through Meme Exposure on Facebook: An Eye-tracking Experiment

Julia Dalibor

Political knowledge is a central concept in understanding the mechanisms that promote political participation. Citizens with high levels of political knowledge engage in behavior that contributes to the well-functioning of democracies: They hold more stable opinions and are more likely to translate them into consistent voting behavior (Kleinberg & Lau, 2019). While prior research has investigated learning processes in the context of digital media platforms in general (e.g., Bode, 2016; Boukes, 2019), the effects of specific digital information types on political learning are still understudied.

One information type which might be particularly promising in capturing the audience’s attention, and thereby easing political learning processes, are memes. Memes are a unique mix of visual, textual, and humoristic elements, which are prevalent in most peoples’ newsfeeds. Not only are memes humoristic, they are also user-generated; two factors that in the past have been associated with high levels of generated attention among audiences (Boukes, 2019; Ye et al., 2011). But does the sum of these content characteristics indeed result in higher attention levels compared to other visual information types? And does this, in turn, lead to higher levels of political knowledge?

This study attempts to answer these questions by conducting an eye-tracking experiment. The technology provides an important advantage over self-assessed attention measures as it allows for an ecologically more valid measure of visual attention. In an attempt to widen our understanding of political learning processes, this study aims to investigate how new forms of political information presentation shape knowledge in the citizenry.

 

Viral Violence: How police brutality and protest violence can influence individual’s affective state, support for political outcomes, and social media behavior. 

Neil Fasching @neilfasching

People have a “negativity bias” when it comes to consuming news content, with individuals putting more weight and attention on negative information (Trussler & Soroka, 2014). Past research found negative news produced much stronger psychophysiological responses than positive news (Soroka & McAdams, 2015). Recently, a form of negative news is spreading not only across main-stream media, but also through social media and interpersonal contact: viral videos and images of police brutality and protest violence. 

This study will help understand the individual level explanations why news – and in particular violence – spreads through social media, as well as investigate whether police and protest violence evokes greater outrage than non-politically-charged violence. It will also probe if ideology is crowding out the effect of violence, as it seems likely that outrage is reduced when violence is against their out-group relative to their in-group. Finally, this project will probe the effect of this violence on the political outcomes, such support for the police, as well as social outcomes, such as willingness to share social media and desire to view similar content.

To investigate these outcomes, a large survey experiment will be run. Custom photos and vignettes will be created, with police brutality and protest violence depicted in the experimental conditions and non-political physical altercations or verbal altercations depicted in the control condition. The photos and vignettes will also manipulate group identity, indicating the violence is against a group they support (in-group condition) or against a group they do not support (out-group condition). 

 

Designing virtual reality experiences as a persuasive tool to promote pro-environmental behavior: The longitudinal effects of prompts added to VR experience

Hana Hegyiova

Climate change is one of the most pressing environmental crises that our society is facing. With rather ‘prolonged’ and statistical nature (Weber, 2006), its intangible consequences often tend to feel distant for people and not solvable by an individual him/herself (Ahn, Bailenson, & Park, 2014). Recently, the potential of using virtual reality (VR) to perceive distinct environmental problems has become widespread (Ahn, Bailenson, & Park, 2014) with attempts to provoke pro-environmental behavior (Bailey et al. 2015). However, most of the studies focused mainly on vivid visualizations (Markowitz, Laha, Perone, Pea, & Bailenson, 2018), embodiment (Bailey et al. 2015), or stimulating hard to experience situations (Chittaro & Zangrando, 2010), expecting attitudinal or behavioral change. 

Such experiences are undoubtedly memorable but could lack information on concrete pro-environmental actions to take. This study will look at VR technology’s persuasive potential by employing persuasive technology (PT) principles through the behavior model for persuasive design (Fogg, 2003; Fogg, 2009). The author will examine how being confronted with a prompt embedded in the VR experience itself could make such VR technology more effective compared to a scenario without a prompt. 

Moreover, most of the existing research is focused on the participant’s self-reported attitudes measured right after the experience. However, there is a possibility that these short-term measures reflect a sort of a ‘wow-effect’ that most people tend to experience after experiencing VR. At the same time, these initial effects may quickly wear off in time. Whether the change of attitudes or behavioral intentions has long-lasting effects will be examined by an experimental design focusing on longitudinal data with two measurements: firstly, after the experiences, and then, one month after the study.

New Project: The Role of News Recommender Systems in High-Choice Information Environments

Claes de Vreese (ASCoR, University of Amsterdam), together with Prof. Frank Esser, Dr. Sina Blassnig (IKMZ, University of Zurich) and Prof. Anikó Hannák (Department of Informatics, University of Zurich), have received a project grant from the Swiss National Research Program “Digital Transformation” (NRP 77). The three-year project will run from 2021 to 2023 and investigate the role of news recommender systems in high-choice information environments.

The main aim of this project is to understand how algorithmic news recommender systems affect the production, perception of, and trust in journalism and news. We further aim to develop evidence-based recommendations on how these systems can be (re)built and (re)programmed to meet various normative requirements, including those promoting enlightened citizenship and a pluralistically informed public. The project is internationally comparative across Switzerland and the Netherlands, problem-oriented, and interdisciplinary, combining theoretical and empirical approaches from the social sciences and computer science.

The project is embedded in the Digital Communication Methods Lab and involves a 3 year post doc in Amsterdam.

Results of the 2nd Semester of DigicomLab Thesis Grants 2019-2020

Three exciting research projects have been conducted as part of (research) master theses written at the University of Amsterdam’s Graduate School of Communication in semester 2 of the 2019-2020 academic year that received funding through the third edition of the Thesis Funding Grants. Alicia Gilbert, Nicolas Mattis, and Vladislav Petkevic have conducted digitally innovative studies, covering the important topics of smartphone-imputed stress, political news quality, and negative campaign effects. You can read about the results of their studies here.

The next round of the Thesis Funding Grants is going online at the end of August 2020, stay tuned for updates!

Three proposals selected for the Digicomlab thesis grants this semester

The Research Priority Area Communication and its Digital Communication Methods Lab are happy to announce that three proposals have been selected for the third edition of the Digicomlab Thesis Funding Grants. These grants provide financial support for theoretically-relevant and digitally innovative (research) master theses written at the University of Amsterdam’s Graduate School of Communication for semester 2 of the 2019-2020 academic year.

When the smartphone takes over: The roles of goal conflict and autonomy appraisals in eliciting digital stress

Alicia Gilbert

In today’s media-saturated environments, the use of mobile digital devices like smartphones is ubiquitous, leading users to be permanently online and connected to others. Both cognitive (i.e. online vigilance) and behavioural components (i.e. communication load, media multitasking) of permanent connectedness can elicit stress (i.e. digital stress), which is detrimental to individuals’ well-being and health. The present research deepens the understanding of underlying mechanisms of digital stress by considering the boundary conditions of goal conflict and autonomy appraisals which have been linked to indicators of permanent connectedness (i.e. availability demands, instant messenger use) in the past. It thus integrates two streams of research: Media stressors identified in the digital stress literature are put in context of intervening factors stemming from self-determination theory and research on media use for basic need satisfaction. With an experience-sampling design spanning seven days, indicators of permanent connectedness and stress are observed in a situational context. Concepts can thus be measured in an application- and context-specific manner with only short time lag between a behaviour and its measurement, reducing biases that have previously skewed measurements of e.g., digital media use. Furthermore, temporal fluctuations of effects can be analysed, which is so far scarce in digital stress research and can guide the creation of digital stress interventions.

 

Is quality journalism going bankrupt? An automated comparison of news quality indicators among political news in German print and online as well as national and regional newspapers

Nicolas Mattis

News outlets’ adherence to normative standards for news quality, such as diversity, impartiality, and comprehensibility (Urban & Schweiger, 2014), is crucial for society, as it determines how well news outlets can contribute to a healthy and deliberative democracy (Strömbäck, 2005). In light of worries about the impact that increased online readership (Burggraaff & Trilling, 2017) and the ongoing commercialisation of news production has on news quality (Jacobi, Kleinen-von Königslöw, & Ruigrok, 2016), it is both timely and relevant to examine if and how adherence to these standards differs depending on modality (online vs. offline) and reach (national vs. regional). To move the field methodologically forward, this thesis will develop a comprehensive framework for automatically assessing news quality in German newspapers and examine potential differences between different types of outlets. This will be done by a) combining existing automated news quality indicators, and b) advancing the automated measurement of impartiality by means of supervised machine learning (SML). In a field that still largely relies on manual content analyses, doing so offers new insights into how and to what extent news quality can be measured in an automated manner. The results will not only provide important insights into the current state of news quality in Germany, but also contribute to the field by addressing important questions about both the opportunities and limitations of automated research methods (Boyd & Crawford, 2012) and by developing an overall framework, as well as classifiers and a training dataset for impartiality that future studies can build on.

 

Emotions as the Impetus of Negative Campaigning Effects. Investigating Voters’ Perception of Campaign Negativity, Voter Turnout, and Vote Intention

Vladislav Petkevic

So far, research on negative campaigning has produced inconsistent findings on the effects of negativity on voter turnout and voter intention. Emerging evidence suggests that these inconsistencies stem, at least in part, from the fact that the effects of campaign negativity are a function not of some absolute value of negativity of a political ad – the usual subject of analysis, but rather of a person’s perception of it. The present study addresses this possibility by analyzing the emotions expressed in the UK public’s online discourse around the 2019 UK general election. Moving beyond the customary (sentiment) analysis of the candidates’ rhetoric, natural language processing and supervised machine learning will be employed to estimate the public’s emotional reactions to the candidates’ political campaigns. The estimates wil utilized in three ways. On the most aggregate level, a time series analysis will be conducted to examine the relationship between daily average values of emotions expressed towards a given party and that party’s position in the polls. Constituency-level data will be used to model a relationship between a given constituency’s electorate’s emotional state on the election date and voter turnout in that constituency. Lastly, sentiment expressed in reaction to individual candidates’ campaigns will be compared to the levels of negativity expressed by the candidates themselves. These analyses will elucidate how the conventional measures of negativity in campaigns compare to the public’s perception of it, and how discrete emotions experienced by the electorate affect the electorate’s voting choices.

Physical distancing is not social distancing – Recent report on smartphone use during COVID-19 outbreak

Increase in smartphone use (duration,
pickups, answered notifications) 01 Feb – 24 March

Together with colleagues from Tilburg University and the Ghent University, DigiCom Lab researcher Jakob Ohme has published a report on citizens’ smartphone use around the Covid-19 outbreak in Belgium. Based on MobileDNA smartphone tracking data from 2778 Flemish adults, the researchers see how communication patterns change in time of crisis. After Belgian lock down measures were in place, the average user used their smartphone 213 minutes per day (an increase of about 45 minutes, or 28%). The number of smartphone pickups remains fairly stable over this period. This means that users constantly turn to their smartphones, but use it longer to access news (43% increase), social media (31 % increase) or calling people (62% increase).

The report shows that smartphones are indispensable means for people to stay informed and they help public authorities to reach out to people. Especially the high increase in mobile messaging shows that in times of crisis, the smartphone for many people prevents that physical distancing results in social distancing.

The full report and associated datasets can be found here.

 

New project: Using Natural Language Processing and Generation to develop and test health-related chatbots

Chatbots have great potential for digital communication: they are always accessible, can do many interactions in parallel and engage in personalised interactions. However, the current generation of chatbots tends to have only a very limited set of responses, produce interactions that can be perceived as meandering, and cannot adopt adequately to the user’s knowledge and needs, as derived for the user’s responses. As a result, people tend to dissociate from chatbots quite rapidly, meaning their potential is presently unfulfilled.

In the present project, we integrate expertise from health communication, health psychology, and digital humanities to create and test chatbots that aim to foster long-term relationships between users and chatbots by having chatbots engage in human-like conversations. We aim to do this by iteratively developing chatbots that have a ’theory-of-mind’ and an interaction memory, for which the chatbot can subsequently engage in varied conversations. We use recent insights from computational linguistics, motivational interviewing, and persuasion and test our chatbots for both sexual health promotion and smoking cessation.

The project is a collaboration between University of Amsterdam, Radboud University, and Tilburg University. From the Digital Communication Methods Lab, dr. Gert-Jan de Bruijn is involved in the project and will supervise two PhD students.

Prof. Claes de Vreese appointed Faculty Professor on Artificial Intelligence

Prof. Claes de Vreese has been appointed Faculty Professor on Artificial Intelligence at the Faculty of Social and Behavioral Sciences at the University of Amsterdam as of September 1. Claes is the director of the RPA Communication and its Digital Communication Methods Lab, and Co-PI at the Research Priority Areas Personalised Communication and Human(e) AI.

As a Faculty Professor, Claes will accelerate his work investigating the societal consequences of Artificial Intelligence, including its effects on decision-making, democratic processes and politics. An important aspect of the role is to continue building upon and expanding collaborations within the faculty, as well as across the University of Amsterdam and with external partners, with a strong interdisciplinarity focus.