We release the validated database with real-world Twitter pro-climate messages (distilled from an original pool of all posts from more than 3600 climate communicators) alongside the preprint. Messages were manually rated and categorized (pre-LLM era ;-) with a taxonomy that we newly developled)
That's amazing, congrats! I hope you had a good start at St Andrews.
The same pattern emerged for different political subgroups (right-leaning people liked climate messages overall less, but they agreed with left-leaning participants on the ranking).
Some degree of cross-cultural consensus on which pro-climate messages were liked best: Crowd-based message ratings showed significant overlap between the U.K., China, the US and Germany.
Also, crowd-based message ratings were predictive for downstream outcomes: they significantly predicted shifts in climate change concern and information search behavior. This makes empirical message ratings useful for assessing the quality of climate communication.
Cheers to my co-authors @tspampatti.bsky.social and Ulf Hahnel.
Check out my interview with PLOS Climate, and while you are at it, the other editions of this cool series on young climate researchers. I particularly enjoy reading what others think about the future of climate research: a mix of interesting ideas with a dose of optimism.
📯 New Preprint out 📣
Honestly, this started out as a side project that I had put on the back burner for a while. Then, I realized I could combine the original data with the data from our climate (dis)information sampling study, and found quite unexpected and cool results: doi.org/10.21203/rs....
We're excited to announce the Spring 2026 Cooperation Colloquia!
With @yanchenl.bsky.social @minhuayan.bsky.social @saptarshi.bsky.social @aronszekely.bsky.social @shafenbraedl.bsky.social @shuxianjin.bsky.social @smconstantino.bsky.social and more
Sign up here: list.ku.dk/postorius/li...
To improve climate communication, it is crucial to understand how different audiences respond to pro-climate messages, particularly in information environments where accurate information competes with...
