Fall Research Expo 2024

Location Analysis of Climate Change Discourse on Twitter

The effects of climate change are becoming more widespread and severe as a result of continuous fossil fuel burning, deforestation, and more. Despite nearly unanimous scientific agreement on anthropogenic climate change, public concern and action remain limited due to misconceptions, highlighting the urgent need for better communication strategies. With this in mind, my PURM project addressed the communication side of the climate crisis, specifically online climate change discourse. Social media provides a good tool for understanding the climate change conversation because it captures an unfiltered, up to date, diverse array of information from the general public. A data set of 28,845 tweets was used for this study to gain greater insight into the state of the conversation. Previous literature has found that there is a relationship between location (i.e., where a person lives) and a person’s opinions of the climate crisis, so my study used location as a tool to expand upon this information. More specifically, location was divided into three distinct variables: state political affiliation, coastal versus inland, and climate change vulnerability index. Each was chosen due to the unique role they play within climate change discourse: state political affiliation was looked at because of the polarized nature of this conversation, coastal versus inland was looked to see if the differing effects of climate change experienced by states (sea level rise and floods for coastal states and drought for inland) would be apparent within the conversation, and, finally, climate change vulnerability index in order to see if tweets from more vulnerable states contained higher levels of concern and worry about the crisis and vice versa for less vulnerable states. I conducted a content analysis on a random subset of 1,000 tweets in order to uncover regional differences in climate change perception and opinions based on tweets posted. I divided these tweets based on state political affiliation (each tweet was assigned “blue” or “red” depending on if the twitter posted the tweet from a state that went red or blue in the 2024 election) to see how the polarized aspects of the climate conversation took form in the online conversation. I chose to conduct the content analysis using state political affiliation because a quantitative analysis of the tweets revealed the most interesting results for this variable. I found that trending topics regarding the climate change conversation included: President Zelenskyy’s climate-focused speech, the Big Oil lawsuit in California, and Hurricane Lee; tweets mentioning “@POTUS” frequently attacked President Biden. Blue states were more likely to acknowledge climate change as a scientifically supported phenomenon (60.56% of tweets) compared to red states (34.63%). Overall, the conversation was predominantly negative across both red and blue states, with blue states showing 53.83% negative sentiment and red states 61.10%––unfortunately, positive sentiments were relatively rare.

PURM - Penn Undergraduate Research Mentoring Program
College of Arts & Sciences 2026
Advised By
Michael Mann
PURM - Penn Undergraduate Research Mentoring Program
College of Arts & Sciences 2026
Advised By
Michael Mann

Comments