The problem seems to be the WVS data, will send an email to their main office

Haven’t checked more countries.

In 2006, Germany had according to F025 not a single protestant respondent. Instead they are found in the “other” category. But are correctly categorized in all other years

In short, the ANES data shows:
📉 Social media use is shrinking
💥 Twitter/X posting has moved ~50 points to the right
🧩 Platforms are splintering
🔊 Fewer people are talking — but those still talking are more politically extreme

Herzlichen Glückwunsch und viel Spaß!

Reminder to myself: look up social trust scores for India and Indonesia tomorrow.

The worrisome problem is how to change higher education programs whose whole point is to train students how to write. If the goal is to learn academic writing, then in-class assignments are strictly speaking „un-aligned“ with the learning outcomes. But all writing at home is now under AI suspicion

How big is the field though