Emotions on the agenda? A Natural Language Processing approach to identifying partisan framing of climate change denialism, fatalism, and solutions in US Congressional speeches

Past studies of agenda setting have shown how emotions of dread and fatalism can influence public opinion, especially in the context of climate change (Merry & Mattingly, 2023). This study uses the theoretical frame of agenda-setting in the US Congressional context to document changes in the emotional valence used in discussing climate change. We measure the use of denialist, fatalist, and solutions-oriented emotional framings of climate change for Republican and Democratic congress-people in Congressional floor speeches between 1987-2016. We use Khodak et al.’s (2018) Natural Language Processing (NLP) methodology—á la carte (ALC) word embedding—as well as other more traditional NLP methods (i.e. dictionary-based classification, and nearest neighbors) to analyze the semantic meaning of climate-change related words in a corpus of floor speeches in the United States House of Representatives and Senate during this period. We compare the semantic meeting–the inherent significance or connotation within a particular context–of the words as used by Democratic and Republican members of Congress as well as their use over time. We find evidence of differences between parties and changes in those differences over time, with Republican congress-people evidencing more denialist and less solutions orientated use of words related to climate change. We find little evidence of differences between parties in their relation to climate fatalism, although there is a shallow increase in this sentiment over time. These findings add to the agenda-setting literature about climate change, documenting the significance of party-specific emotional strategies in shaping political agendas and providing a potential theoretical connection between emotion-based strategies of agenda-setting actors and public opinion.