Modeling complex associations between socio-demographic factors and climate policies using Conditional Inference Trees
摘要
Public attitudes toward climate change in the United States are strongly related to political identity, but also shaped by socio-demographic factors that interact in complex ways. This study uses Conditional Inference Trees (CIT) to examine how political identity and socio-demographic characteristics, including education, income, age, and ethnicity, are associated with climate skepticism and support for specific climate and energy policies, using Gallup Poll Social Series survey data from U.S. adults collected between 2000 and 2020. The models show that party identification is the factor most strongly associated with climate change perception, and they show how age, education, income, and ethnicity are associated with differences within parties. Extending the analysis from simple climate beliefs to 14 finer-grained policy proposals, we examine how support for these policies varies both across and within partisan groups. Age is strongly associated with Republicans’ policy preferences: younger Republicans tend to support renewable energy development and oppose coal development, even as they remain opposed to carbon taxes, whereas education is the factor most strongly associated with differences in Democrats’ views on regulatory and tax-based policies. We also identify policy domains with broad support across parties, including government spending on solar energy and the development of alternative fuels, alongside proposals that meet broad cross-party opposition, such as legal limits on household energy use.