-- KamilaKoronska - 16 Jul 2024

Mapping Uncharted Waters: The Presence and Influence of Post-Truth Spaces in European Mainstream media

Team members: Sofia Bonfanti, Liliana Bounegru, Fabio Daniele, Michaela Dubóczi, Kamila Koronska, Elena Negrea-Busuioc, Richard Rogers, Giulia Tucci, Ahmad Zareie

1. Introduction

‘Post-truth’ is considered a ‘condition’ of an information network (or analytically demarcated actor-source set) rather than a ‘period’ where previously there was more truth or more modernity on offer. It has been described as a network condition whereby ‘“alternative facts”replace actual facts, and feelings have more weight than evidence’ (McIntyre, 2018).

This project seeks to locate post-truth spaces in Facebook networks in the Czech, Polish, Romanian and Slovak languages, found through keyword queries concerning the Russia-Ukraine war, July, 2023-July, 2024. These countries are selected given their targeting by Russia’s “influence and disinformation operations” (Dodge, 2022) as well as the work conducted there by fact-checkers and journalists seeking to expose them. In fact, this project contributes to methodologies that surface leads for fact checkers. It also seeks to measure the impact of post-truth spaces through an influence metric that gauges the extent of their interpenetration with more mainstream and central media spaces (where the constitution of the mainstream is also of interest).

2. Research Questions

1. To what extent are post-truth spaces present as well as impactful in Facebook networks concerning the Russia-Ukraine war in Eastern Europe?

2. How do the presence and positioning of these spaces compare across the Czech, Polish, Romanian and Slovak languages?

3. Methodology

A) Data collection

We queried CrowdTangle by Meta for Facebook posts on pages and public groups related to the Russia-Ukraine war in Czech, Polish, Romanian as well as Slovak using both stance-taking as well as more neutral terms for the period of July 2023 to July 2024. The post types sought include statuses, links and YouTube videos. The countries (or languages) were chosen given their targeting by Russia’s “influence and disinformation operations” (Dodge, 2022) as well as fact checking activities in them that seek to expose them. It also enables a comparison between the languages concerning the presence of these spaces as well as their impact.

B) Cosine similarity

From a bipartite network of Facebooks groups and channels (sources) pointing to the URLs shared within them (targets), we generated a monopartite network of URLs similarity based on the calculation of cosine similarity. Steps followed to calculate the cosine similarity between the URLs:

1. Plot a URL sharing Matrix, where each unique URL represents a row and each FB page / group, a column.

2. Normalize the matrix through dividing the value in each cell (the n. of times the URL was shared in a page/group) by the sum of all values in the same row.

3. Calculate the cosine similarities between every pair of URLs (e.g. cosine similarity between ULRN and URLN+1, on the next row and so forth).

4. The similarities among the connected URLs are represented by the weight of their edges.

C) Data analysis and methodology

The approach is a network analysis of accounts (pages and public groups) and the URLs they post. The bi-partite network is then projected as a mono-partite network (only URLs) where the edges indicate URLs shared by the same sources. The networks are optimised through a visual network analysis with the goal of accentuating the clustering (and the structural holes between them). The clusters are labelled, where the point of departure are such typical cluster categories as these (this is based on previous work, quoted in Reference section):

a) post-truth - disinformation and anonymous sources, Russian/Belarussian state media content, Russian / Belarussian propaganda sources

b) populist - includes borderline content ('awful but lawful')

c) nationalist / conservative - an older establishment perhaps recently reactionary

d) mainstream media where the question concerns what constitutes that

e) pro-EU / pro-NATO / pro-Ukraine

f) historical disputes - sources that comment on modern and contemporary history, especially in the context of military campaigns in 20th and 21st century

The cluster types may differ per country.

4. Findings

Generally we found a pattern of cluster types across countries, where there are distinctive clusters we labelled as Russian propaganda, borderline content (often populist), nationalist/conservative, mainstream media, pro-Ukraine / pro-EU as well as historical disputes. The populist and the nationalist/conservative occasionally overlap. In certain countries, there are additional clusters; in Polish language space, for example, there is a political vendetta space, where the sources concern the investigation of the former ruling government. The post-truth space per country is rather prominent and central, though its influence differs quite significantly.

Romanian

The Facebook Romanian language space surrounding the war in Ukraine is characterised by several distinct clusters. The Mainstream Media space includes international mainstream media outlets. Historical Disputes focus on anti-communist and anti-USSR historical content within Romanian and Moldovan online spaces. The Pro-Ukraine Pro-EU area consists of Moldovan anti-Russia social media and a broader multi-platform space that supports Ukraine and opposes Russia. The Nationalist Conservative segment is a nationalist, spiritual, anti-elitist social media space with a pro-Russia inclination. The Populist Borderline area encompasses anti-Western, conspiratorial, far-right web and social media, along with pro-Russia propaganda within the Romanian online sphere. Post-Truth includes Russian disinformation and propaganda web spaces (including Russian state media), Russian institutional websites disseminating pro-Russia content (e.g. the Russian embassy), and pro-Russia propaganda in the Romanian online space. Lastly, the Place Name space includes Romanian and international tourism websites that were queried due to 'keyword pollution.'

Polish

The Facebook Polish language space discussing the Russian invasaion on Ukraine could be break down into the following clusters: The Mainstream Media space includes daily newspapers, major news portals, news in Ukrainian, finance news, tabloids, and journalistic investigations. Historical Disputes focus on Volhynia historical archives, hagiographies of the doomed soldiers, the anti-communist Polish resistance movement, Polish anti-Semitism, and the Eastern Borderlands. The Nationalist Conservative segment consists of Law and Justice official accounts, conservatives, and Christian free-market supporters. The Place Name space includes Facebook events. Post-Truth includes Belarusian propaganda, TikTok ultranationalists, and Ria Novosti copy-posters disseminating pro-Russia content. We have also observed the presence of a Political Vendetta cluster, where Civic Platform politicians, their paid influencers (such as Pablo Morales), and outlets dedicated to archiving and researching the wrongdoings of the Law and Justice Party government reside.

Czech

The Facebook Czech language space discussing the conflict in Ukraine encompasses the following communities: The Mainstream Media cluster includes both domestic mainstream media and foreign sources, such as Ukrainian websites and neutral sources informing about security issues. The Pro-Ukraine Pro-EU cluster features commentators and journalists, fundraising campaigns, mainstream media, fact-checkers, and pro-Ukraine commentators, along with neutral sources. The Populist Borderline cluster is characterized by populist and nationalistic politicians, including SPD and Aliance národních sil. The Slovak Post-Truth cluster comprises Russian media, pro-Russian Facebook profiles, pro-Russian and nationalistic sources, and various conspiracy and fringe websites.

Slovak

In Slovak, the most prominent discussions about the war in Ukraine on Facebook can be characterised by the following clusters of media sources: The Mainstream Media cluster consists of mainstream media mixed with pro-Ukraine oriented individuals on Facebook. The Pro-Ukraine Pro-EU cluster features pro-Ukraine sources and fundraising campaigns. The Nationalist Conservative cluster includes nationalist politicians and anti-NATO and anti-Western sources. The Place Name cluster involves Facebook events focused on spirituality and "soul healing," along with popular YouTube channels featuring individuals, celebrities, bands, and singers. The Czech Post-Truth cluster comprises pro-Russian disinformation media, Russian domains and sources, pro-Russian Telegram channels, and foreign (including Russian) domains.

5. Discussion

The prominence and visibility of post-truth spaces, which we define as clusters characterised by problematic content (such as social media accounts of ultranationalist groups), difficult-to-verify information, or Russian and Belarusian state and propaganda media, vary between the language spaces we have analysed.

In the Romanian and Czech language spaces, post-truth spaces are more prominent, albeit at the periphery of the graph. In contrast, in the Polish language space, they are visibly smaller but dangerously integrated into the centre of the graph. Meanwhile, in the Slovak language space, they occupy a significant portion of the space and are also at the centre of the network.

This method allows us to visually assess the integration of these sources in the respective language spaces on Facebook and to comment on possible avenues of integration. For example, in the Polish space, potential avenues for integrating post-truth narratives could come from sources classified as Historical Disputes and Nationalistic/Conservative social media accounts or media outlets. In the Slovakian language space, post-truth spaces touch both conservative clusters and significantly smaller mainstream media clusters, which means they could be potentially next in line to be infiltrated.

Knowing how and where problematic sources are strategically placed is key to determining how to counterbalance these with strategically chosen FIMI countermeasures. One example could be initiating marketing strategies to disrupt the Historical Disputes cluster in the Polish space with high-quality content related to the subject matter.

This method also serves as a lead generator, grouping sources based on sharing behaviour similarity. For each country, we were able to produce a lengthy list of problematic/disinformation/Russian state-sponsored sources, some of which were previously unknown to fact-checkers and are clearly in the early stages of building their visibility.

6. References

1. Fabio Giglietto, Nicola Righetti, Luca Rossi, and Giada Marino. 2020. Coordinated Link Sharing Behavior as a Signal to Surface Sources of Problematic Information on Facebook. In the International Conference on Social Media and Society (SMSociety'20). Association for Computing Machinery, New York, NY, USA, 85–91. https://doi.org/10.1145/3400806.3400817

2. Santo Fortunato, Community Detection in Graphs (2010) https://arxiv.org/abs/0906.0612

3. Dodge, M. (2022) Russia’s Influence Operations in the Czech Republic, Poland, and Romania, Washington, DC: National Institute for Public Policy.

4. McIntyre, L. (2018). Post-truth. Cambridge, MA: MIT Press.

Previous work:

Multi-layered Structure of Narratives relating to Ukrainian Refugees on Polish Telegram

Mapping Facebook post-truth spaces in Eastern European countries, Sweden and the Netherlands

Mapping Facebook post-truth spaces in Eastern European countries - Moldova