ABSTRACT
To analyze the factors affecting US public concern about the threat of climate change between January 2002 and December 2013, data from 74 separate surveys are used to construct quarterly measures of public concern over global climate change. Five factors should account for changes in levels of concern: extreme weather events, public access to accurate scientific information, media coverage, elite cues, and movement/countermovement advocacy. Structural equation modeling indicates that elite cues, movement advocacy efforts, weather, and structural economic factors influence the level of public concern about climate change. While media coverage exerts an important influence, it is itself largely a function of elite cues and economic factors. Promulgation to the public of scientific information on climate change has no effect. Information-based science advocacy has had only a minor effect on public concern, while political mobilization by elites and advocacy groups is critical in influencing climate change concern.
Disclosure statement
No potential conflict of interest was reported by the authors.
Notes
1. While the sample size is significantly larger than those in previous analyses, it is still insufficient to reach robust conclusions regarding causality. However, given the limitations of continuous time-series data on public opinion regarding climate change, these limitations will not be overcome for several years if only a single nation is utilized in the analysis. What we hope to provide is a series of initial and tentative empirical conclusions that can contribute to a fuller effort to expand the scope of the analysis to develop a transnational analysis of the drivers of aggregate levels of climate change concern.
2. See Huxster et al. (Citation2014) and Brulle et al. (Citation2012) for an extended discussion of this approach and a list of the questions used in constructing this index.
3. Some scholars have advanced concerns about Stimson’s algorithm (e.g., see McGann Citation2014) on the grounds that it is theoretically ad hoc because it does not sufficiently link individual-level response behavior to the observed aggregate outcomes in the algorithm. While we recognize this ongoing debate, there is a sizable body of literature that has used the Stimson algorithm to produce aggregate measures of policy mood and many of these studies have provided ample evidence of its validity. Thus, we side with the bulk of the literature and employ the Stimson approach.
4. The tone scores are calculated based on a detailed coding scheme of all climate-related editorials and press releases published between 2001 and 2013. Each item was given a score of ‘0’ if the statement was ‘Anti-Climate Change Science or Climate Change Action’ or a ‘1’ if the statement reflected support for ‘Pro-Climate Change Science or Climate Change Action’. Data were drawn from the Lexis-Nexis Congressional database. Climate-related press releases were identified using the following keywords: ‘Climate Change,’ Global Warming,’ ‘Greenhouse Effect,’ and ‘Carbon Dioxide.’
5. Note that the model is likely over-fitted given that we have only 52 cases. While simulations to assess the accuracy of SEM models (e.g., MacCallum et al. Citation1996; Wolf et al. Citation2013) suggest that our sample size is sufficient, results become biased as the number of IVs grows. These studies suggest that the primary consequence of having few cases and many IVs is inflated standard errors, which would make statistical significance less likely to be achieved. Given this, the results from our fully specified model should be considered with caution. Findings from our models with more limited specifications ( and ) serve as robustness tests.
6. We considered alternative proxies to capture the dissemination of scientific information in academic publications. Specifically, we assessed whether or not the number of climate-related articles in Nature could account for shifts in media coverage on this issue. Regardless of how we operationalized scientific information on climate change, it did not significantly influence media coverage on this issue.