with @jniemannlenz, @mattrevers, @LasseRodeck.

The AfD’s success on TikTok (in our data, generating more engagement than other parties combined) is linked to strategic messaging: it combines divisive content with relatable fears concerning inflation and security, indicating a shift in right-wing populist tactics online.

We argue that this may be a deliberate choice during the election run-up, as our analysis showed that typical out-group attacks and culture war themes generated the highest engagement rates. Yet, the AfD's content often prioritizes addressing voters' concerns.

Despite its typical anti-elite rhetoric, our findings show that the AfD strongly focuses on voters' concerns, such as financial anxieties and everyday fears, which significantly drive high engagement rates.

With rightwing populism rising globally and unexpectedly early German elections, our new working paper feels timely: Beyond Anti-Elitism & Out-Group Attacks: How Concerns Shape the AfD’s Populist Representation on German TikTok During the 2024 EU Elections

RT @kathaesa: How does #polarisation manifest in media & communication at breaking points? Our new @icsjournal article identifies 5 key sym….

RT @AvishaiGreen: Had a great time presenting work I did about automated hypocrisy accusations detection with @Hendrik__Meyer @_paugc @myrt….

We highlight that different #LLMs have varying strengths in hypocrisy accusation detection. Political hypocrisy accusations are harder to detect than personal ones. More context, and therefore larger models, tend to perform better at detecting these nuanced accusations.

The idea for this paper originated a year ago at the @hackingcommsci of #ICA23. We used the Reddit European Sustainability Initiatives corpus and identified a clear gap in current #NLProc for detecting hypocrisy accusations.

Now out in #KONVENS proceedings: @AvishaiGreen & @myrthereuver presented our (@_paugc,@ankekat1000,@_Y_X_Y_) paper in Vienna: "A Few Hypocrites: Few-Shot Learning and Subtype Definitions for Detecting Hypocrisy Accusations in Online Climate Change Debates"

RT @scheufele: 63 algorithmic interventions designed to make #Facebook temporarily provide less polarizing and more reliable content likely….

Today at #ecrea24, @louprsl presented our analyses on intermedia-agenda setting processes between social media and journalism in debates concerning disruptive climate protest -- great ending of the conference! With @MichaelBruegge

RT @anamariadutceac: Using #ECREA2024 to soft launch our Volkswagen Stiftung project studying the relationship between media and polarizati….

with @louprsl, @hwessler, Helena Rauxloh, Mike Farjam.

. and our #ECREA2024 panel on (de-)polarization just finished, with a lot of interesting presentations and discussions!.@ChristelvanEck, @mattrevers, @gelovanishota, @MichaelBruegge did such a great job illustrating very different perspectives on the matter

With @patrick_zerrer and @Kudusch.

This morning at #ECREA2024,.@HannahMarieBu presented our multi-platform research (content & network analysis) on advocacy coalitions (and antagonists) of climate protesters on Telegram, Mastodon, and Twitter/X. Great panel with interesting discussions on Cross-Platform analysis!

. And last but not least @MichaelBruegge & @anamariadutceac.

With a lot of great researchers, including.@ChristelvanEck @ashleymuddiman @gelovanishota @hwessler @mattrevers @lonenerup, Mike Farjam, Helena Rauxloh.