Interventions intended to reduce beliefs in climate change had much stronger effect than interventions to increase beliefs; we might tend to cling to information that gives us hope, downplaying the consequences

Rode, Jacob B., Amy Dent, Caitlin N. Benedict, Daniel B. Brosnahan, Ramona L. Martinez, and Peter Ditto. 2020. “Influencing Climate Change Attitudes: A Systematic Review and Meta-analysis.” OSF Preprints. August 20. doi:10.31219/osf.io/bjen5

Abstract: Researchers interested in climate change communication have investigated how people respond to messages about it. Through meta-analysis, the current research synthesizes the multitude of experimental studies on this topic to uncover which interventions are most effective at influencing attitudes about climate change. The meta-analysis focuses on experimental studies that included a control condition and measured climate change attitudes among participants in the United States. After a large literature search, 396 effect sizes were retrieved from 76 independent experiments (N = 76,054 participants). Intervention had a small, significant positive effect on attitudes, g = 0.08, 95% CI [0.06, 0.10], p < .001. Surprisingly, type of intervention was not a statistically significant moderator of this effect, nor was political affiliation. However, type of attitude was a significant moderator: the treatment-control difference in attitudes was smaller for policy support than for belief in climate change, indicating that policy attitudes are more resistant to influence than belief in climate change. Other moderators and publication bias were also tested. We conclude with policy implications and recommendations for future research.


Policy Support is Difficult to Influence

Our results indicate that climate change belief is much easier to influence than support for climate change policy. Unfortunately, policy support is arguably more important than belief with belief often seen as only instrumentally important to drive support for climate policies. Even if interventions were not effective for beliefs, meaningfully moving the dial in policy support would produce important implications for policymakers. For example, P. S. Hart and Feldman (2018) found that people were more receptive to policy when it was framed around air pollution rather than climate change, suggesting that there may be ways to garner policy support among climate skeptics without changing their minds about the existence of climate change. Although it is difficult to sway policy attitudes, there may be ways to influence policy support without first changing belief in climate change. Targeted interventions for specific policies may be articularly effective for meaningful climate action (e.g., highlighting policy effectiveness; Reynolds et al., 2020).


Future Directions

Future research should examine how and why attitudes about climate change are more sensitive to negative than positive messages about it. Moreover, future research could examine ways of offsetting this increased malleability to skeptical messages about climate change. Doing so would reduce the potential for increased uncertainty around climate science that so often sparks skepticism (e.g., Dunlap & Jacques, 2013).
In addition, our findings indicate that interventions were more effective if they were conducted in samples with a higher percentage of participants identifying as female and that attitude phrasing interacted with political affiliation. While we did not develop predictions about gender or attitude phrasing, the results may spark interest in future primary research on the topics. For example, a substantial body of research has investigated gender differences in environmental concern (e.g., Bloodhart & Swim, 2020; McCright, 2010). Future research could continue to investigate not only gender differences in climate change beliefs but also in differential response to interventions. In addition, we found that interventions were slightly more effective for conservatives when they used the term “global warming” than “climate change” (with the opposite pattern for liberals and moderates). These results add to a growing body of work on this topic (e.g., Soutter & Mõttus, 2020) and pose a new way of studying responses to the terms, namely comparing if interventions are differentially effective between them.
Our findings also suggest that research should investigate ways of making climate policy palatable. For example, previous research shows that avoiding the term “tax” is useful for garnering policy support (Hardisty et al., 2010, 2019). Additionally, framing policy as being supported by the ingroup may increase support, although findings using this strategy are somewhat mixed (Bolsen et al., 2019b; Fielding et al., 2020; Zhou, 2016). Future research should continue to focus on policy support as an intervention outcome (for a review, see Kyselá et al., 2019) and consider testing different types of policy support (e.g., word framing) along with the effectiveness of different types of interventions (e.g., ingroup support) on these more nuanced aspects of policy support.
Finally, the current meta-analysis is one of the first to organize the varied interventions on climate change attitudes into specific categories. While other unnamed categories of interventions may remain, future research could build upon past work in the same category identified in this project to help facilitate future attempts to integrate and reconcile this growing -- and potentially fracturing -- area of research.