Public perception bias of air pollution risk based on semantic analysis and decoupling theory

Air pollution poses serious challenges to both the environment and human health. Influenced by individual sensory characteristics （such as visual sense） and environmental context （such as visibility）， the public often struggles to accurately assess air pollution risk levels. By integrating textual semantic analysis with decoupling theory， this study proposed a novel Public Air Pollution Risk Perception Bias Index to quantify the gap between subjective risk perception and actual risk levels， taking Chengdu as a case study for empirical analysis. Results showed that： 1） In the years with more frequent air pollution warnings and the autumn and winter seasons with more severe air pollution， the number of air pollution-related texts peaked， indicating heightened public concern during these periods. 2） There are seasonal differences in the public’s subjective risk perception scores of air pollution， with average values of -1.5 in autumn and winter， and 4.9 in spring and summer. This indicates that the public subjectively perceives air pollution to pose greater health risks during autumn and winter seasons. 3） A discrepancy existed between the public’s subjective risk perception and the actual risk levels， with both overestimation and underestimation occurring. In autumn and winter， due to the more direct and intense sensory impact of pollutants， which mainly consisted of particulate matter such as PM2.5 and PM10， the proportion of overestimating the risk （26.9%） was significantly higher than that in spring and summer （15.1%）. 4） On the one hand， regulatory authorities should develop differentiated pollutant control strategies based on the seasonal characteristics of pollution； On the other hand， they should strengthen public awareness and accurate understanding of air pollution risks through information dissemination and public education.