By ChatGPT: AfD state election scenario - drivers, counterforces and consequences

The content of this page is licensed under the Creative Commons Attribution 3.0 License.
Quick view of the model (open complete model with iMODELER):

present thumbnailKai Neumann (#1) has provided a description of the model with the iMODELER Presenter.

Description

Here's my chat with chatGPT:

Can you create an iMODELER model (.imm file) with a qualitative cause-and-effect model on the question of the right wing AFD party winning upcoming state elections. Explore what leads to it, what might prevent it, and what it could mean. Within the qualitative model you can weight the impact, e.g. with 10 for weak, 20 for medium, and 30 for a strong impact. Attached an example file that also uses weighting of connections.

Done. I created the .imm model using your example file as the structural template.

Download the iMODELER model – AfD state election scenario (.imm)

It contains 42 factors and 55 weighted connections, consistently using 10 = weak, 20 = medium, 30 = strong. I separated three parts of the model visually: drivers of an AfD electoral victory, factors that could prevent/reduce it, and possible consequences.

Importantly, the central factor is “AfD becomes largest party in state election”, with a separate downstream factor “AfD leads or participates in state government.” That distinction is particularly relevant to the current elections: Mecklenburg-Vorpommern and Berlin vote on 20 September 2026, and current polling in Mecklenburg-Vorpommern has the AfD around 36% and clearly in first place, while coalition arithmetic could nevertheless keep it out of government. 

Among the strongest drivers in the model are dissatisfaction with government, perceived lack of control over migration, protest/change voting, issue ownership on migration/security, low institutional trust, and weakness/fragmentation of competing parties. Counterforces include credible solutions on migration/security, satisfaction with government, economic improvement, trusted competing candidates, turnout among non-AfD voters, strategic voting, concern about AfD policies, and AfD campaign mistakes.

I also built in some feedback loops, for example electoral success → greater organizational resources/normalization → greater credibility in later elections, and electoral success → counter-mobilization → higher turnout among opponents. That should make the Insight Matrix / impact analysis in iMODELER considerably more interesting than a simple tree. 



Here's chatGPT's description: Qualitative scenario model of conditions that can increase or decrease the likelihood that AfD becomes the largest party in an upcoming German state election, and of possible consequences. Weights: 10 = weak, 20 = medium, 30 = strong. The model is exploratory and not a forecast or endorsement. It distinguishes electoral plurality from actually leading or joining a state government.

Add Your Comment:

Comments (1)

Kai Neumann

Kai Neumann

This is a 10 minute exercise - imagine what could be modeled with a more precise prompt. And then we look at it, believe it and stop thinking for ourselves - or we take some time and use it to sharpen our thinking?
Reply