More than scientists: How message and messenger attributes influence viewers’ climate change intentions

PLOS One | https://doi.org/10.1371/journal.pone.0331672 September 10, 2025 4 / 20 science of climate change, the IPCC report contained more cautious and formal language, fewer emotional and concrete words, and more complex syntactic structures than the NIPCC report, and the NIPCC authors used emotional language to challenge and discredit the IPCC [38]. Furthermore, Bailey et al. [39] found that newspaper articles tended to adopt the IPCC’s qualifying and hedging language (e.g., likely, possible, probable, may) when describing climate science. Scientists often use more cautious and less explicit language because their claims are inferential and future research may necessitate reinterpretation. But in the context of climate change, this may give the lay public the impression that their predictions are mainly guesswork, and therefore do not present a great risk or threat [30]. Perceived lack of cer- tainty and risk may be countered by impressions of authority and power, and these characteristics have also been implicated by the results of linguistic frame analyses of climate communications [40–42]. For example, Benites-Lazaro et al. [40] found that language used in 35 videos and multimedia presentations produced by Brazil’s sugarcane industry portrayed ethanol as a “green hero” that can help to reduce greenhouse gases and save the planet, and Landrum et al. [42] found that the 2015 papal encyclical contained a large amount of words that indicated authority, expertise, and confidence of the author.

The predictive potential of linguistic frames is exemplified by experimental studies that have demonstrated different responses to equivalent linguistic expressions, such as climate change versus global warming. For example, Whitmarsh [43] found that individuals tend to associate climate change with natural causes and global warming with human causes, and that the term global warming is associated with greater concern, greater certainty, and stronger belief that individual behaviors can make a difference. Other linguistic studies have indicated that the effects of these two expressions are moderated by partisanship. For example, Villar and Krosnick [44] found that conservatives who answered questions about climate change (as opposed to global warming) considered the climate issue to be more serious, whereas liberals viewed the issue as more serious when asked about global warming. Similarly, [45,46] found that linguistically framing survey questions as global warming elicited more skeptical responses among conservatives, but both terms produced similar effects on liberals’ climate change beliefs, and that unseasonal cold weather decreased belief in global warming, but not climate change, among climate sceptics. In a recent study, Findlater et al., [47] found that linguistic framing of weather and climate risks by South African grain farmers strongly predicted their adoption of climate-resilient agricultural practices (i.e., conservation agriculture). High adopters of conservation agriculture used more rational (agricultural, economic, and cognitive) language when describ-

ing weather-related issues than low adopters. However, farmers’ rational language only weakly predicted their adoption of conservation agriculture, which was instead primarily determined by less use of survival and emotional language. Although textual corpuses of climate change communications have been linguistically analyzed, and the predictive effects of language have been experimentally examined, few (if any) previous studies have combined both approaches in climate- change research. Nevertheless, researchers that have used the LIWC software to analyze rationality, emotional tone, achievement and affiliation motives, and other linguistic styles and motivations in digital messages in other subject domains have made some relevant observations. For example, a study by Mackrill et al. [48] analyzing the linguistic elements of TED talks found that authentic styles, use of ‘I’ and social, positive words were given more emotional ratings. Whereas, analytic styles, fewer use of ‘I’ and positive emotions were rated more educational. They also found that talks by academics (as compared to non-academics) ranked higher on analytic style and clout, and were rated as more fascinating, informative, and persuasive. Less analytical, more positive, and more authentic TED talks were also found to be more popular, while clout had no significant effect on popularity [49]. Other studies found that the interactive effect of authenticity and clout in Olympian’s tweets positively influenced its retweets [50] and advertisements containing analytical thinking language is

more engaging in online pet adoption [51] and is more persuasive [52]. Other studies [53] analyzing motives in text using the drive dimensions of LIWC categorizations, found that affiliation and achievement motives in online user reviews were found to be less useful, while power reviews were perceived to be