Abstracts

Jean-Philipe Cointet- Sciences Po Paris

The opinion of people on different issues of public debate has been traditionally studied with polls and surveys. More recently, network ideological scaling methods have shown that digital traces in social media platforms can be used to mine opinions at massive scales. With this class of methods, social media users are embedded in a latent ideological space. In this talk, I will discuss how the dimensions of this emergent space relate to classical attitudes in political science: trade protectionism, immigration, European integration, etc. Such variables measured through traditional surveys in various countries can be propagated to Twitter users to estimate their distribution on social network. I will further illustrate the interest of such embeddings beyond opinion estimation for media studies (fake news propagation) and collective action analysis (yellow vests movement),

Tom Willaert- Vrije Universiteit Brussel

In our present society, news websites play a pivotal role in documenting as well as shaping societal debates. However, along with other innovations such as social media, news websites have been affected by the rapid expansion of the digital realm, which has led to information overload, fragmentation and polarization. As more and more voices and perspectives enter the public domain through different (social) media, journalists and editors face the challenge of presenting information to their readers in a way that does justice to the many aspects of societal debates, and that maximizes the content and opinion diversity of online news. Over the past years, researchers at the VUB AI Lab, in close collaboration with other European partners, have worked on a range of methods, instruments and tools called media observatories that might help journalists gain insight into online streams of information. This talk will present and demonstrate the most recent outcomes of these efforts, with the aim of opening up a discussion about future directions for this line of research: how can we for instance operationalize digital methods to fight disinformation or other forms of information pollution? And which societal as well as technical challenges should be prioritized?

Massimo Airoldi, Emlyon Business School – airoldi@em-lyon.com

Culture is now a largely datafied and machine-readable matter. Platformized social discourses, symbolic struggles, value judgments, collective identities, social interactions, consumption practices and experiences leave digital traces ordinarily exploited by surveillance capitalism. Academic researchers can use the byproducts of this ubiquitous extractive process to fruitfully understand socio-cultural processes at an unprecedented scale. However, doing cultural sociology with digital traces also presents important epistemological, ethical, methodological, and practical challenges. These will be discussed through examples derived from my YouTube-based research on musical taste and cultural distinction (Airoldi 2021a; 2021b; 2017; Airoldi, Beraldo and Gandini 2016). Particular attention will be given to how platform-based forms of algorithmic confounding have substantially turned cultural phenomena into techno-social outcomes, and the (major) research implications of this.

Marloes Geboers - Universiteit Van Amsterdam

The platformed circulation of social media images adds all kinds of complexities to the analysis of meaning. To begin with, the processes of circulation are not merely user-driven but also algorithmically defined. Social media images are ephemeral, reworkable and less ‘stable’. This fluidity allows people to connect with societal issues through, for example, altering iconic news images or through aligning with issues by sharing and endorsing imagery. Captions, comments, tags and engagement metrics are but some examples of data points that provide clues for the interpretation of meaning. I demonstrate protocols that blend digital methods techniques of data collection and sampling with qualitative visual analyses. Methodological challenges often arise from the standardizing nature of platforms. Engagement metrics and other metadata provide a simplified glimpse of sociality that emerges from platform strategies of quantification. Through combining automated data collection, quantitative methods for sampling, and qualitative assessments of networked visual content, I aim to ‘unstandardize’ –or recontextualize– the quantitative traces of engagement with images. Considerations over query design, its affordances and constraints for datasets, research questions as well as research scope will be discussed. Challenges specific to working with image data will also be addressed. These lie in retaining connections between the multimodal elements of visual social media content. One way to go about this, is to repurpose machine vision to ‘decontextualize’ image content within demarcated discursive frameworks. This pertains to new methodological implications which are discussed as well.

Alessandro Caliandro - University of Pavia - alessandro.caliandro@unipv.it

Despite growing interest, there is a shortage of research about the methods and challenges that concern researching ephemeral digital content. In this presentation I would like to discuss two research strategies to study Instagram Stories. Instagram Stories allow users to share moments of their everyday lives in a documentary and narrative style; their peculiar feature is ephemerality, as each Story lasts for 24 hours. The presentation (a) explores how to bypass the Instagram API closure and (b) engages in an attempt at ‘circumventing the object of study’, taking advantage of how individual users archive Instagram Stories on other platforms (here, YouTube). In so doing, I would like to contribute to the debate that seeks to innovate and ‘repurpose’ digital methods in a post-API environment. The presentation will also cover the ethical issues and challenges implicated in the study of Instagram stories. Lastly, the presentation will reflect on the methodological possibilities and advantages of cross-platform analysis, discussing also its possible future developments.

Roser Beneito-Montagu -Cardiff University

This presentation will address methodological questions about the study of mediated social relations in everyday life using digital and ‘traditional’ data and tools and perspectives of qualitative research. The qualitative inquiry has recognised that offline and online social interactions are entangled and to capture everyday life and, accordingly, researchers need to employ research designs that consider various kinds of data and different qualities of digital data because using exclusively digital data limits the observability of a social phenomenon in this milieu. Some of the difficulties in studying mediated relationships as another layer of everyday life are: (1) how to link online observations with offline life and vice versa, and (2) how to analyse the broad array of multimodal data generated from these trans-situated observations. This presentation will provide a detailed account of one way to address these questions while studying everyday life practices on social media using digital data to grasp people’s experiences, feelings and emotions considering their material “doings”. Digital practices are both relational and affective. Yet, these practices are situated in particular topographical sites and material places. In turn, these practices do things to people and their environments too. Specifically, I will introduce a research design to study affectivity and social media use in later life that employs digital data and traditional data. Through the application of mixing-methods, it will show how to capture -in a complex and multi-layered way- the digital practices of older people as situated activities. Rather than questioning traditional data (such as interviews or fieldnotes), this approach advocates a combination of traditional and digital data to understanding current everyday life.