Machine learning models find that about 60% of Americans are confused about climate risks and support for climate policies
Understanding public support for climate policy is critical to shaping effective strategies to reduce the impact of climate change. However, predictive policy support has long been challenging because many factors can affect public opinion. A team of researchers led by Professor Asim Zia of the University of Vermont, including Professor Katherine Lacasse of the University of Rhode Island, Professor Nina Feverman and Professor Louis Gross of the University of Tennessee, Brian Beckage of the University of Vermont, and Brian Beckage of the University of Vermont, and Brian Beckage of the University of Vermont, have developed a new machine learning approach to better understand the complexities of these complexities. Their research, published in the Sustainability of Peer Review Journals, introduces an approach called structural equation model, a statistical approach that explores the relationship between different factors by considering mutually inconsistent probability and uncertainty, which helps analyze different factors such as people’s concerns about climate change, their beliefs, political views, politics, competition, beliefs about them, competition, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them, support for them,
Unlike older methods that rely on assumptions about which factors are most important, this new approach uses machine learning, an artificial intelligence that allows computers to find patterns in data and improve predictions without explicit programming to find patterns in large sets of data. “By using unsupervised machine learning technology, we let the data itself show us the connection between different factors, eliminating bias from human speculation,” Professor Zia explained. The study used data from a long-term survey called Climate Change in American Thought, which spanned over a decade and included responses from representatives nationwide. This new method has higher accuracy predictions than traditional statistical methods.
One of the most surprising findings of the study was the discovery of a previously unacceptable “lukeless supporter” who made up most of the U.S. population. Unlike strong supporters or staunch opponents of climate policy, these people are confused about the climate risks and ambivalence of supporting or opposing climate policy actions. Research shows that people don’t think about climate risks in one way. Instead, the study divided risk perception into two types: analysis (logical assessment) and affective response. Professor Zia noted: “We found that emotions such as worry play a greater role in shaping policy support rather than purely logical assessments of climate risks.” In addition, he noted: “Both emotional and analytical news can convince 60% of people to be confused, mostly mild, contradictory public to support collective action.”
Research by Professor Zia and colleagues also confirms that political views and beliefs about climate science strongly influence policy support. People who trust scientific consensus, based on a large amount of evidence, general agreements between experts on climate change are more likely to support climate policy, and those who do not tend to oppose them. However, machine learning models show that political identity is a connection between a person and certain political beliefs or the political parties that shape their viewpoints on the problem and does not completely determine people’s views. By considering factors such as risk perception, race and demographic context, the model has a deeper understanding of the responses of different groups to climate policy.
These findings are of great significance to policy makers and those working to increase public support for climate action. Understanding different categories of policy supporters can provide more effective communication strategies. For example, attracting lukewarm supporters, through emotional connections, rather than focusing solely on scientific facts, may be more effective. The study also highlights the need to include public opinion trends in climate policy planning, ensuring that policies reflect attitudes that have changed over time.
By using machine learning, the study provides a new way to understand what drives public support for climate policy. It provides a data-based approach to address one of the biggest challenges in climate exchange: reducing political divisions and encouraging a broader consensus on the need for climate action.
Journal Reference
Zia, A., Lacasse, K., Fefferman, NH, Gross, LJ, and Beckage, B. Sustainability, 2024, 16, 10292. doi: https://doi.org/10.3390/su162310292
About the Author
Asim ZiaResearch, teaching and outreach activities focus on promoting the sustainability and resilience of integrated social environment systems. Asim Zia serves as a professor of public policy and computer science in the Department of Community Development and Applied Economics and has made a secondary appointment in the Department of Computer Science at the University of Vermont (UVM). He is the director of the Institute for Environmental Diplomacy and Security (IEDS) and Ph.D. University of Vermont Sustainability Policy, Economics and Governance Program.
Katherine LacasseHe is a professor of psychology at Rhode Island College. Her research focuses on risk perception and behavioral change, applicable to climate change, local ecosystems, environmental infrastructure projects and health behaviors. Much of her recent work has been done as part of an interdisciplinary team, focusing on incorporating feedback from human social systems into climate and epidemiological models.
Professor Nina Feverman The research focuses on epidemiology, evolutionary and behavioral ecology, and mathematical self-organizing behavior, especially systems of network description. Although Fefferman Lab’s research often focuses on diseases in human and/or animal populations, and how behavioral ecology associated with diseases and diseases affects short-term survival and long-term evolutionary success of a population, people in the lab are involved in making computer networks that make social behavior safe, making social behavior in grassroots organizations vulnerable to radicalization.
Louis J. Gross He is an emeritus professor of ecology and evolutionary biology and mathematics at the University of Tennessee Knoxville. He is honorary director of the National Institute of Mathematics and Biosynthesis, a NSF-funded center to foster research and education at the interface between mathematics and biology. He is a member of the American Association for Scientific Development, the American Society of Ecology and the Society of Mathematical Biology.
Professor Brian BeckageVery interested in computing and complexity. He has a specific interest in climate change, species diversity, forest dynamics, and the intersection of social processes and natural systems. He emphasized the use of quantitative methods to study these systems, including statistical, analytical and computational models.
