The story built slowly through August and peaked on election night, 6 September, when three of every four articles in the corpus were about Sachsen-Anhalt. That is more than twice the Berlin election two weeks later. The AfD winning the speaker's chair on 6 October drew a much smaller second wave.
Every headline naming Sachsen-Anhalt, Siegmund or Magdeburg's parliament, coded by its main frame with the same codebook as the Berlin piece. Choose a frame to follow it.
Through the campaign the AfD is mostly a contender: polls, duels, events. After the vote the main question is who governs. Extremism stays a steady minority frame at about one headline in five and never leads. In the final week, the one stretch where politicians, economists and churches warned loudest, it reaches its high of one in four, with the most negative tone.
Two September elections, two parties at the edge of the spectrum. On both measures, the AfD in Sachsen-Anhalt was described in extremism terms about half as often as the Berlin Left. The Verfassungsschutz classes the AfD's Sachsen-Anhalt branch as confirmed right-wing extremist.
This measures what outlets wrote, not what either party is. One plausible reading: the AfD's classification is old news, carried as a label rather than a story, while the Left's antisemitism row was new. The other: a party that is likely to govern gets covered on its odds of governing. The data can't separate the two.
The Berlin piece also showed a hand-coded sample of 47 national AfD headlines at 38% extremism framing. The larger vocabulary measure here does not reproduce that (16% for the AfD outside Sachsen-Anhalt), so treat that small sample with caution.
Headlines whose core is the cordon: who will cooperate with the AfD, tolerate it, block it or walk out on it. After the AfD took the speaker's chair with four more votes than it and the BSW have, more than half the coverage turned to whether CDU members had helped, and to the CDU's plan to leave the chamber when Siegmund stands.
Civil society here means economists who modelled an AfD government (DIW, IWH), business groups (IHK, BDI, the hotel association), churches, the Jewish community and queer groups. In the Berlin story the criticism came mostly from would-be coalition partners and, at the end, from the party itself. The AfD shows no internal critic in any period.
The taz framed close to half its Sachsen-Anhalt headlines as extremism. The Tagesschau framed none of its 28 that way: it covered polls, results, voter flows and government formation.
| Finding | Test | p |
|---|---|---|
| Frame mix changes over the five periods | χ², df 24 | 4 × 10⁻⁶ |
| Less extremism framing than the Berlin Left (headlines) | Fisher | 0.002 |
| Less extremism vocabulary than the Berlin Left (articles) | Fisher | 1 × 10⁻⁵ |
| Cordon dominates after the speaker vote | Fisher | 8 × 10⁻⁵ |
| AfD in Sachsen-Anhalt differs from the AfD elsewhere | Fisher | 1.0 · no |
Data. Politikradar collects the top articles from the RSS feeds of German online outlets every day, about 85 a day. Each article is stored as headline plus a short AI-written summary; the full text is not kept.
Story. Headlines naming Sachsen-Anhalt, Siegmund, Magdeburg or Sven Schulze from 1 June to 7 November 2026. 298 headlines, 17 excluded as unrelated (football, the Magdeburg attack trial, a hospital scandal), 1 exact duplicate dropped.
Coding. Main frame, valence towards the AfD, who carries negative framing, and whether the headline is about the cordon. The codebook is the Berlin one with one extra flag for the cordon.
Limits. The AI summaries get details wrong in places: several give the wrong election year, one names the wrong chancellor, seat counts differ between summaries. Nothing here depends on those details, and the election result is given only as “about 44%”, where the summaries agree.