The differential impact of climate interventions along the political divide in 60 countries

Climate change mitigation faces a significant hurdle: the political polarization of beliefs surrounding the issue. A substantial body of research demonstrates a strong correlation between political ideology and climate change beliefs globally. Individuals identifying as liberal consistently exhibit higher levels of belief in and support for climate action compared to those identifying as conservative. This polarization negatively impacts climate action and policy support. This study investigates this polarization in beliefs and behaviors, and tests whether various behavioral climate interventions differentially affect beliefs, policy support, and individual actions across the liberal-conservative divide worldwide. Two competing hypotheses are explored: belief-behavior polarization congruence (believers act, skeptics do not) and belief-behavior polarization incongruence (beliefs and actions are not uniformly aligned across the political spectrum). The latter hypothesis suggests that factors beyond belief, such as cognitive biases, perceived control, personal costs, and social norms, can influence pro-environmental behavior. This study also aims to assess the effectiveness of various climate interventions across the ideological spectrum, considering the possibility that certain interventions may be more successful with liberals than conservatives. Previous research indicates that strategies like scientific consensus messaging might be less effective on conservatives, even potentially causing reactance. Therefore, the research examines whether diverse climate interventions yield differential results for liberals and conservatives.

Extensive research documents the global link between political ideology and belief in climate change. Meta-analyses reveal that political ideology is the strongest predictor of climate change belief, with liberals showing greater acceptance than conservatives. This pattern holds across various countries, including Australia, the UK, and the US. Studies highlight that making political identity salient can causally influence climate change beliefs and policy support. The question arises whether this belief polarization translates into behavioral polarization. While beliefs often predict behavior, factors like cognitive biases, perceived control, and personal costs can moderate this relationship. In the context of climate change, this is known as the 'green gap', representing the incongruence between pro-climate beliefs and sustainable behaviors. Although often attributed to liberals, this gap has also been observed in conservatives, as demonstrated by farmers adopting pro-environmental practices despite climate change skepticism. The effectiveness of various climate change interventions also varies across political divides; studies show differential responses to scientific consensus messaging and moral framing depending on political leanings.

This international collaboration study involved a large-scale experiment conducted in 60 countries with 51,224 participants. Eleven climate interventions, crowdsourced and vetted by behavioral science experts, were compared to a control group. The interventions were diverse, including those that aimed to address political polarization directly. Participants were recruited predominantly online, with convenience and snowball sampling methods used. No pre-determined sample size was set. After random assignment to a condition, participants completed measures of climate change beliefs (four statements rated on a 0-100 scale), climate policy support (nine statements rated on a 0-100 scale), and an incentivized behavioral task: the Work for Environmental Protection Task (WEPT). The WEPT involved completing pages of numerical tasks, with each completed page resulting in the planting of a real tree through donations. Participants also provided demographic information, including their political ideology (on a 0-100 scale, 0 being liberal and 100 conservative). Data analyses involved linear and ordinal mixed-effects models to account for non-independence within countries and participants. Bayes factors were used to quantify evidence for null hypotheses. Robust and weighted least squares mixed models were conducted to address potential violations of distributional assumptions.

The study replicated the global polarization of climate change beliefs and policy support, with liberals exhibiting stronger beliefs and greater support than conservatives. However, there was no significant difference in tree-planting behavior between the two groups. Analysis suggested that this lack of behavioral polarization was due to conservatives engaging in tree planting despite lower belief levels, rather than liberals failing to act on their beliefs. Specifically, the analyses indicated a conservative-oriented green gap. Three interventions consistently boosted climate beliefs and policy support across the ideological spectrum: emphasizing effective collective actions, writing a letter to a future generation, and writing a letter from the future self. One intervention (emphasizing scientific consensus) increased climate action among liberals only; no intervention significantly increased climate action among conservatives. In fact, several interventions even backfired among conservatives, decreasing their tree-planting efforts. The study found that, on average, self-identified conservatives' climate beliefs predicted their behaviors less than those of self-identified liberals, further supporting the conservative-oriented green gap interpretation.

The findings challenge the simplistic notion that the green gap results solely from liberals' inaction. The data suggest a more complex picture, where conservatives act despite skepticism, possibly due to motivations beyond climate concern, such as preserving traditional values of nature or seeing tree planting as an alternative to broader policy action. This has significant implications for designing interventions; strategies need not always focus on belief change to elicit action. The varying effectiveness of interventions across ideologies emphasizes the need for targeted approaches. While some interventions worked across the board (improving collective action messaging, future-self projections), others, like scientific consensus messaging, were effective only for liberals. Norm-based interventions didn't generally improve outcomes, possibly due to the sample's cultural diversity, highlighting the need for culturally sensitive interventions and potentially challenging the generalizability of theories from WEIRD (Western, Educated, Industrialized, Rich, Democratic) countries.

This global study reveals a complex relationship between political ideology and climate action, suggesting that conservative-oriented green gaps also exist. Several interventions effectively enhanced climate beliefs and support across ideologies, while scientific consensus messaging proved effective only for liberals. Future research should explore interventions that promote climate action without requiring belief change and focus on developing culturally sensitive, targeted interventions. The data and analysis code are publicly available to encourage further investigation and theory refinement.

The study's operationalization of climate action (WEPT) as a single, individual-focused behavior may limit the generalizability of findings to collective or systemic actions. The attrition rate (36.4%) in the online study should also be considered when interpreting results. Future research should examine broader behaviors and implement strategies in more controlled settings with lower attrition. The interventions were homogeneously applied, neglecting the potential of targeted, ideologically tailored strategies shown to be more effective in other research.