idw – Informationsdienst Wissenschaft
Nachrichten, Termine, Experten
Nachrichten, Termine, Experten
Social media platforms provide a rich source of data for academic study. A research group at Saarland University has now analysed the Twitter accounts of some 8,600 politicians from twelve countries. The team, led by Ingmar Weber, who holds an Alexander von Humboldt Professorship in AI at Saarland University, wanted to know whether between 2021 and 2023 Twitter prioritized tweets from politicians from the right of the political spectrum. The short answer for the period in question is no. In June 2023, however, Elon Musk blocked free access to Twitter data for academic research purposes.
Sometimes research work ends just when things are starting to get exciting. Someone who knows what that feels like is graduate computer scientist Brahmani Nutakki, who is working on her PhD under Professor Ingmar Weber at the Institute for Societal Computing at Saarland University. Over a period of two years, Brahmani Nutakki and the two Saarbrücken-based political scientists Rosa M. Navarrete and Giuseppe Carteny analysed the tweets from 6,550 Twitter accounts, which were associated with some 8,600 politicians from twelve countries. ‘Sadly, our analyses came to an abrupt end in June 2023 when Elon Musk, who acquired Twitter in October 2022, removed free academic access to Twitter’s data,’ explained Brahmani Nutakki. Nevertheless, Nutakki was able to draw some important conclusions from the data that the team had been able to collect up to that point. The analysis has now been published in the respected free-to-access scholarly journal ‘Journal of Quantitative Description: Digital Media’.
The researchers selected twelve countries with the highest user activity on the Twitter (now X) platform, namely Argentina, Brazil, Canada, Colombia, France, Germany, India, Japan, Mexico, Spain, the United Kingdom, and the United States. For each of these countries, the researchers divided the parliamentary representatives into two groups depending on whether they could be assigned more to the political left or more to the political right. ‘We opted for this binary distinction for the sake of simplicity and because the use of the labels “left” and “right” to denote particular political positions is broadly accepted around the world. To determine how politicians and parties should be assigned within this binary classification scheme, we consulted numerous sources such as expert surveys and party manifestos,’ explained political scientist Rosa M. Navarrete.
Having analysed several thousand tweets for each politician, the researchers found no significant difference between the dissemination of content from left-leaning or right-leaning politicians across all countries studied. ‘In the period we studied from 2021 to 2023, we did, however, observe a considerable increase in the number of likes, a slight increase in the likes-per-retweet ratio and a significant reduction in the number of retweets post Musk acquisition. We believe that this is because early followers of a Twitter account are more politically engaged and tend to support content more actively. But as the political message gets spread to a broader audience, it becomes more likely that the number of likes will go up, while the retweet rate declines,’ said Brahmani Nutakki. The researchers used the metric ‘likes’ rather than ‘views’, as the latter were only introduced after Elon Musk acquired the platform and were not available for previous tweets. They were, nevertheless, able to establish a correlation between the two metrics.
Nutakki also faced a number of problems when analysing the huge amounts of data involved. Many of these difficulties were related to the dynamic changes occurring on the Twitter platform between November 2022 and June 2023. One such change was the transformation of Twitter Blue from a premium subscription service for high-profile accounts with high levels of user engagement into a low-cost subscription service that only required a verified phone number. This new feature was then temporarily removed shortly after it was introduced. A similar fate befell the rules prohibiting hateful conduct, which were intended to prevent attacks on individuals and were also removed after Musk’s takeover. ‘Changes to Twitter’s functionality also resulted in changes in the user population, which made it difficult for us to attribute the changes we observed to a single specific cause,’ said Brahmani Nutakki.
‘We believe that our study is the first to provide a global analysis of how trends in political engagement have shifted over time by examining how user activity on Twitter has evolved during the recent changes that the platform has undergone. Our findings indicate that political content on the platform reached a wider audience following these changes. In the period before June 2023, we found no evidence of a significant difference in engagement patterns between users on the left and those on the right of the political spectrum,’ explained Professor Ingmar Weber. Professor Weber also highlighted an article in the Washington Post last October that reported that those Republican politicians most active on X (formerly Twitter) were posting more frequently, getting more views and had more followers than top-tweeting Democrats, though no evidence of platform censorship was found. Other observers have pointed out that since Musk’s takeover, the Twitter/X platform has become problematic by failing to remove reported hate tweets and by reactivating many previously suspended accounts.
‘These worrying changes to social media platforms such as Twitter/X are jeopardizing the democratic structures in many countries and require continuous scrutiny by academic researchers. We are calling for the implementation of legislation such as Article 40 of the Digital Services Act so that academic researchers can regain access to social network data. It is simply not acceptable that access to these vast amounts of data is either unaffordable or provided so selectively that no meaningful analyses can be conducted,’ said Ingmar Weber.
Background: Interdisciplinary Institute for Societal Computing (I2SC)
I2SC advances interdisciplinary research at Saarland University and provides a platform for interaction and knowledge sharing between the social sciences, the humanities and computer sciences. The institute is headed by Ingmar Weber, Alexander von Humboldt Professor for Artificial Intelligence, and Daniela Braun, Professor for Political Science. The research conducted at I2SC covers two core areas: Computing of Society and Computing for Society. The first of these areas concerns the use of computational methods to understand societal phenomena. The second relates to the use of these methods to facilitate digitally assisted interventions that improve society.
Prof. Dr. Ingmar Weber
Tel: +49 681 302-70788
E-Mail: iweber@cs.uni-saarland.de
Dr. Rosa Navarrete
Scientific Researcher
Tel: +49 681 302-2368
E-Mail: rosa.navarrete@uni-saarland.de
Brahmani Nutakki
Scientific Researcher
Tel: +49 681 302-70789
E-Mail: bnutakki@cs.uni-saarland.de
‘Is there anything Left? A Global Analysis on Changes in Engagement with Political Content on Twitter in the Musk Era’, Brahmani Nutakki, Rosa M. Navarrete, Giuseppe Carteny and Ingmar Weber, Interdisciplinary Institute for Societal Computing (I2SC), Saarland University
DOI: https://doi.org/10.51685/jqd.2025.004
http://www.i2sc.net
https://journalqd.org/article/view/8875/7391 - Journal of Quantitative Description
Ingmar Weber, Humboldt Professor of Artificial Intelligence
Thorsten Mohr
Universität des Saarlandes
Merkmale dieser Pressemitteilung:
Journalisten, Wirtschaftsvertreter, Wissenschaftler
Gesellschaft, Informationstechnik, Medien- und Kommunikationswissenschaften, Politik, Wirtschaft
überregional
Forschungsergebnisse
Englisch
Ingmar Weber, Humboldt Professor of Artificial Intelligence
Thorsten Mohr
Universität des Saarlandes
Sie können Suchbegriffe mit und, oder und / oder nicht verknüpfen, z. B. Philo nicht logie.
Verknüpfungen können Sie mit Klammern voneinander trennen, z. B. (Philo nicht logie) oder (Psycho und logie).
Zusammenhängende Worte werden als Wortgruppe gesucht, wenn Sie sie in Anführungsstriche setzen, z. B. „Bundesrepublik Deutschland“.
Die Erweiterte Suche können Sie auch nutzen, ohne Suchbegriffe einzugeben. Sie orientiert sich dann an den Kriterien, die Sie ausgewählt haben (z. B. nach dem Land oder dem Sachgebiet).
Haben Sie in einer Kategorie kein Kriterium ausgewählt, wird die gesamte Kategorie durchsucht (z.B. alle Sachgebiete oder alle Länder).
