Mapping climate discourse to climate opinion: An approach for augmenting surveys with social media to enhance understandings of climate opinion in the United States

RESEARCH ARTICLE Mapping climate discourse to climate opinion: An approach for augmenting surveys with social media to enhance understandings of climate opinion in the United States Jackson BennettID1,2, Benjamin RachunokID1, Roger Flage3, Roshanak Nateghi1,2 1 School of Industrial Engineering, Purdue University, West Lafayette, Indiana, United States of America, 2 Ecological Sciences and Engineering, Purdue University, West Lafayette, Indiana, United States of America, 3 Department of Safety, Economics, and Planning, University of Stavanger, Stavanger, Norway * benne105@purdue.edu Abstract Surveys are commonly used to quantify public opinions of climate change and to inform sus- tainability policies. However, conducting large-scale population-based surveys is often a dif- ficult task due to time and resource constraints. This paper outlines a machine learning framework—grounded in statistical learning theory and natural language processing—to augment climate change opinion surveys with social media data. The proposed framework maps social media discourse to climate opinion surveys, allowing for discerning the region- ally distinct topics and themes that contribute to climate opinions. The analysis reveals sig- nificant regional variation in the emergent social media topics associated with climate opinions. Furthermore, significant correlation is identified between social media discourse and climate attitude. However, the dependencies between topic discussion and climate opinion are not always intuitive and often require augmenting the analysis with a topic’s

most frequent n-grams and most representative tweets to effectively interpret the relation- ship. Finally, the paper concludes with a discussion of how these results can be used in the policy framing process to quickly and effectively understand constituents’ opinions on critical issues. Introduction Enacting sustainable climate policy requires understanding public opinions on climate change. This is complicated by geographic variations in climate change belief in the United States, where county-level belief in climate change ranges from 43-80% of the population [1]. These variations have been attributed to endogenous factors such as regional heterogeneity in racial and ethnic composition [1], media consumption [2], and the beliefs of an immediate social PLOS ONE PLOS ONE | https://doi.org/10.1371/journal.pone.0245319 January 14, 2021 1 / 16 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 OPEN ACCESS Citation: Bennett J, Rachunok B, Flage R, Nateghi R (2021) Mapping climate discourse to climate opinion: An approach for augmenting surveys with social media to enhance understandings of climate opinion in the United States. PLoS ONE 16(1): e0245319. https://doi.org/10.1371/journal. pone.0245319 Editor: Natalia Grabar, STL UMR8163 CNRS, FRANCE Received: February 21, 2020 Accepted: December 28, 2020 Published: January 14, 2021 Copyright: © 2021 Bennett et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original

author and source are credited. Data Availability Statement: All files relevant for this study are available at the public Github repository: https://github.rcac.purdue.edu/ benne105/climate_twitter_data Note that for Twitter data, only tweet IDs (rather than full tweets) are published, as it is against Twitter’s privacy policies to publish full tweets. However, full tweet information can easily be retrieved using the Twitter API (brief instructions on repository page).