Yale Program on Climate Change Communication

We are pleased to announce the publication of a new article, “Field interventions for climate change mitigation behaviors: A second-order meta-analysis” in the journal Proceedings of the National Academy of Sciences.

To mitigate climate change, it is important to know which interventions are most effective, and which behaviors are most changeable. Field experiments conducted in the real-world are one of the best ways to assess behavior change. In this study, the research team conducted a comprehensive meta-analysis of prior field experiments that attempted to increase climate change mitigation behavior, which included 430 primary studies from 10 previous meta-analyses.  

The study found that mitigation interventions are generally effective (d _unadjusted = 0.31, 95% CI [0.30, 0.32]). The overall average effect size was about 12 percentage points. The most effective type of interventions used social comparison—such as highlighting the behaviors or attitudes of others compared to the person receiving the message. Financial incentives were also particularly effective.

The study also examined a range of behaviors and found, for example, that interventions were more effective at changing conservation behaviors (e.g., reducing electricity use) than transportation behaviors (e.g., reducing driving). These results highlight the need to make strategic decisions about whether it is better to target behaviors that are high-impact but harder to influence, or behaviors that are lower impact but easier to influence. We hope this study is useful for researchers and practitioners seeking to promote pro-climate behaviors.

It is important to note that meta-analyses like these have limitations. Namely, results from meta-analyses are often too optimistic because positive results are more likely to be published in the scientific literature than failed interventions (e.g., publication bias). The research team conducted a wide range of tests and adjustments to understand the role of these factors. Adjusted estimates of the average effects of field interventions ranged from as large as 7 percentage points to as small as 2 percentage points. These wide-ranging estimates might be driven by the analyses containing widely different interventions: some are light-touch large-scale interventions (e.g., sending people a home energy report), and some are strong small-scale interventions that are more direct and include face-to-face interactions.

You can find more detail in the full research article, which includes the adjustments for publication bias, as well as additional interpretation and sensitivity analyses.