🌍 Why it matters
Our findings show that negative campaigning in multiparty systems is not random mudslinging. It’s a functional response to ideological crowding, shaped by issue salience and competition.

#PolCom #Elections #DigitalPolitics #CSS #NLP

📌 Key Findings
1️⃣ On salient, contested issues (economy, ecology) → higher spatial pressure = more negativity
2️⃣ On low-salience or depoliticized issues (culture, COVID-19) → higher spatial pressure = less negativity
3️⃣ Negativity is thus a strategic tool to stand out, when the issue context allows it

🛠️ Methods & Data
- 166,402 tweets 🐦 from 1,228 party accounts (7 Bundestag parties)
- NLP pipeline: Named Entity Recognition, stance detection (XLM-RoBERTa), issue dictionary
- A new measure of “spatial pressure” capturing issue-specific ideological crowding

🔍 What we explore
How parties used negative campaigning on specific issues during the 2021 German federal election – and how crowded ideological space (“spatial pressure”) shapes when parties go negative.

6/6 🌍 Why It Matters
How pivotal issues like COVID-19 are (de-)politicized on social media impacts the quality of democratic discourse in digital spaces.

Our findings stress the need to tackle echo chambers&foster critical, inclusive debates online.

#PolCom #COVID19 #BTW25 #EchoChambers #CSS #NLP

5/6 📌 Key Findings (cont’d):
2️⃣ Depoliticization:
Political actors largely avoided COVID-19 in electoral debates, sidelining critical public health discussions.

3️⃣ Selective politicization:
Health policy leaders like Lauterbach (SPD) & Spahn (CDU/CSU) couldn’t avoid the issue due to their roles.

4/6 📌 Key Findings:
1️⃣ Polarization via echo chambers:
Two groups dominated:
• “Safety-first” (strict measures)
• “Freedom-first” (anti-restrictions).

COVID-19 became a battleground for ideological divides.

3/6 📊 Methods & Data
Analyzed X/Twitter data:
• 7.37M posts mentioning the election.
• 3.19M posts about COVID-19.

We calculated Echo Chamber Scores (ECS) based on retweet networks and studied political actors’ engagement in COVID-19 debates.

2/6 🔍 What we explore:
How COVID-19 was discussed—or avoided—on X/Twitter during the 2021 German federal election campaign.

The pandemic was polarizing, debated within ideological echo chambers. Yet, many political actors strategically depoliticized the topic to avoid alienating voters.

Explore #EUDigiParty and download it here: dataverse.harvard.edu/dataset.xhtm...

🧵2/3: Designed for researchers in #polcomm, #partypolitics, & #CSS), #EUDigiParty supports in-depth analysis of digital party campaigns across Europe. It includes handles for #Facebook, #Instagram, #TikTok, #X / #Twitter, and #YouTube, plus Partyfacts-IDs for easy linking to other datasets.

I can offer you ze german bubble
@ulrikeklinger.bsky.social
@sophiahunger.bsky.social
@jasminriedl.bsky.social
@wiebkedrews.bsky.social
@sebstier.bsky.social
@mikecowburn.bsky.social
@corinnaoschatz.bsky.social
@favstats.eu
@benguinaudeau.bsky.social
and myself ;)

Wir analysieren alle (Re-)Tweets der Kandidierenden von @AfD_Bayern, @Gruene_Bayern, @fdpbay, @BayernSPD, @fw_bayern und @CSU für ein Mandat im Landtag. 🗳️📊🐦
Unsere Echtzeitanalysen finden Sie hier: dtecbw.de/sparta
#ltwby23 #Bayernwahl2023 #Bayernwahl (4/4)