Characterizing the semantic features of climate change misinformation on Chinese social media.

Climate change misinformation leads to significant adverse impacts and has become a global concern. Identifying misinformation and investigating its characteristics are of great importance to counteract misinformation. Therefore, this study aims to characterize the semantic features (frames and authority references) of climate change misinformation in the context of Chinese social media. Posts concerning climate change were collected from Weibo between January 2010 and December 2020. First, veracity, frames, and authority references were manually labeled. Then, we applied logistic regression to examine the relationship between information veracity and semantic features. The results revealed that posts concerning environmental and health impact and science and technology were more likely to be misinformation. Moreover, posts referencing non-specific authority sources are more likely to be misinformed than posts making no references to any authority references. This study provides a theoretical understanding of the semantic characteristics of climate change misinformation and practical suggestions for combating them.