Globally representative evidence on the actual and perceived support for climate action

The study addresses the human behavioral dimension of climate action, asking to what extent individuals worldwide are willing to contribute to the global common good of climate stability and how they perceive others’ willingness to contribute (WTC). Recognizing that cooperation is costly and subject to free-rider incentives, the authors focus on four cooperation-related aspects: individual willingness to incur personal costs for climate action, approval of pro-climate social norms, demand for political action, and beliefs about others’ support. They examine global patterns and determinants of WTC, particularly across countries differing in income, temperature, and vulnerability, and analyze whether beliefs about others’ WTC are calibrated or systematically biased. A central hypothesis is that conditional cooperation links individuals’ own WTC to their beliefs about others, and that systematic underestimation (pluralistic ignorance) may hinder cooperation.

The paper situates its contribution within the behavioral science and commons governance literature. Foundational work on common goods and free-riding (Hardin) and institutional governance (Ostrom; Dietz et al.; Besley & Persson) underscore the need for cooperation and policy. Research on social norms (Cialdini et al.; Bicchieri; Nyborg et al.; Fehr & Schurtenberger; Constantino et al.) highlights the role of injunctive norms in promoting cooperative behavior through internalization and enforcement. Evidence on conditional cooperation (Fischbacher et al.; Fehr & Fischbacher; Gächter; Rustagi et al.) suggests individuals contribute more when they believe others will. The study also draws on literature on pluralistic ignorance and misperceived social norms (Allport; Geiger & Swim; Mildenberger & Tingley; Bursztyn et al.; Sparkman et al.), which shows that underestimating others’ support can impede action. Broader work on climate perceptions and policy attitudes (Capstick et al.; Eom et al.; Lee et al.; Dechezleprêtre et al.) provides context, while IPCC assessments and related economic-climate studies motivate the urgency of action. The authors extend this literature by offering globally representative evidence linking actual support, perceived support, and cooperation mechanisms.

Design: The Global Climate Change Survey was fielded within the Gallup World Poll 2021/2022 across 125 countries that account for ~92% of the global population, 96% of GDP, and 96% of GHG emissions. The nationally representative samples cover residents aged 15+. The global sample totals 129,902 individuals, with most country samples ≈1,000. Sampling and modes: Probability-based sampling with stratification; interviews conducted via computer-assisted telephone interviewing (common in high-income countries) or face-to-face (common in low-income countries); at least 3 callbacks for face-to-face and 5 for telephone. Telephone used random-digit dialing or national phone lists; face-to-face used clustered PSUs and random-route household selection with random respondent selection. Median response rates: 65% (face-to-face) and 9% (telephone). Weights correct for unequal selection probabilities and nonresponse; population-adjusted weights used for global aggregates. Measurement: - Willingness to contribute (WTC): Respondents asked if they would contribute 1% of monthly household income to fight global warming (yes/no). If no, asked if willing to contribute a smaller amount (yes/no/none). Responses classified as: willing to contribute ≥1%; willing to contribute >0% and <1%; not willing to contribute. The 1% benchmark aligns with plausible mitigation cost estimates. - Beliefs about others’ WTC: After informing respondents that many others in their country are being asked the same WTC question, respondents estimate how many out of 100 would be willing to contribute ≥1%. - Pro-climate social norms: Whether people in the country should try to fight global warming (injunctive norm; yes/no). - Demand for political action: Whether the national government should do more to fight global warming (yes/no). Not fielded in Myanmar, Saudi Arabia, and UAE. Translation and pretesting: Surveys translated using a multi-step TRAPD procedure (translation, review, adjudication, pre-testing, documentation) to ensure cross-cultural comparability. An initial global pre-test (2020) with cognitive interviews in Colombia, Egypt, India, Indonesia, Kenya, and Ukraine informed refinements, including use of the term “global warming.” Weighting: Gallup base weights adjust for selection probabilities; post-stratification aligns to population statistics by age, gender, education, and region. For global estimates, population-adjusted weights scale country samples by their share of the world population aged ≥15. Validation: The WTC measure predicts incentivized climate donations and self-reported pro-climate behaviors in a representative US sample: those willing to contribute 1% donate 43% more to a climate charity (P<0.001; N=1,993) and are 21–39 percentage points more likely to engage in climate-friendly behaviors (P<0.001; N=1,996). General measures of norms and policy demand correlate strongly with specific norms and policies in a US sample (r≈0.35–0.59; P<0.001). A US online survey produced similar levels of WTC, norm approval, and policy demand as the US Gallup sample. Data integration: Country-level analyses relate WTC to GDP per capita (PPP, 2010–2019), annual average temperature (2010–2019), and an IPCC AR6-based vulnerability index, among other controls (economic, political, cultural, geographic). Subnational analyses (North America) relate WTC to regional temperatures. Regressions include robustness with continent fixed effects and alternative covariates. Implementation issues: Minor errors occurred in Kyrgyzstan (n=4 Uzbek-language interviews missing a sentence) and Mongolia (Q1 asked about <1% contribution; corrected via approximation). Sensitivity checks excluding Mongolia yield similar results.

- Global support: 69% of respondents worldwide are willing to contribute 1% of monthly household income to fight global warming; an additional 6% are willing to contribute a smaller amount; 26% would not contribute any amount. In 114 of 125 countries, ≥50% are willing to contribute 1%; in 81 countries, ≥66% are willing. - Social norms and policy demand: 86% say people in their country should try to fight global warming; in 119 of 125 countries, support exceeds two-thirds. 89% say their national government should do more; in over half of countries, this exceeds 90%. - Cross-country variation in WTC: WTC varies from 30% to 93% across countries. It is negatively correlated with log GDP per capita (p = −0.47; 95% CI [−0.60, −0.32]; P<0.001; N=125) and positively correlated with annual average temperature (p = 0.35; 95% CI [0.18, 0.49]; P<0.001; N=125). A composite IPCC AR6 vulnerability index also positively predicts WTC. Results hold in multivariate regressions with continent fixed effects and additional controls; subnationally (North America), higher average temperature associates with higher WTC. - Conditional cooperation: Country-level WTC is strongly correlated with perceived WTC of others (p = 0.73; 95% CI [0.64, 0.81]; P<0.001; N=125). Individually, a 1-percentage-point increase in perceived WTC is associated with a 0.46-percentage-point increase in one’s own probability of contributing (95% CI [0.41, 0.50]; P<0.001; N=111,134). - Systematic misperceptions (pluralistic ignorance): Globally, there is a 26-percentage-point gap between actual WTC (69%) and average perceived WTC (43%) (95% CI [25.6, 26.0]; P<0.001; N=125). In all 125 countries, average perceived WTC is lower than actual WTC (significantly so in all but one). 81% of respondents underestimate their country’s actual WTC; 73% underestimate by more than 10 percentage points. The perception gap is larger in hotter countries and smaller in richer countries, robust to controls. - Norms and policy coherence: Variation in norms and policy demand is smaller than for WTC; higher temperature predicts stronger norms and greater demand for political action, with no significant relationship to GDP. Across countries, WTC, pro-climate norms, policy support, and beliefs are positively correlated, suggesting reinforcing dynamics.

The findings indicate substantial global willingness to take personal action on climate change alongside overwhelming approval of pro-climate norms and strong demand for greater government action. These dimensions of cooperation are interlinked: consensus on norms likely reinforces WTC and policy support, and enacted policies may in turn strengthen norms. The strong alignment between individuals’ WTC and their beliefs about others’ WTC demonstrates the importance of conditional cooperation. However, widespread pluralistic ignorance—systematic underestimation of others’ support—poses a barrier to collective action by depressing individuals’ own willingness to contribute. Potential drivers of misperceptions include media dynamics that amplify minority skepticism and interest-group campaigns. Given the measured 26-point global perception gap and the estimated 0.46 elasticity of own WTC to perceived WTC, correcting misperceptions could yield sizable increases in cooperative behavior. Communicating the true extent of public support and expectations for governmental action can help trigger positive feedbacks and possibly social tipping dynamics that accelerate climate action. The results can inform socio-climate models that integrate human behavior with earth systems by quantifying how beliefs and norms shape cooperation.

This study provides globally representative evidence that most people are willing to incur personal costs to fight climate change, broadly endorse pro-climate social norms, and demand more government action. It also documents a large and pervasive underestimation of others’ willingness to act, which—coupled with conditional cooperation—can inhibit collective action. Highlighting the true breadth of public support emerges as a scalable lever to enhance participation and policy momentum. The work contributes comparable, behaviorally grounded metrics across 125 countries, links support to vulnerability, and quantifies belief-behavior relationships. Future research could test targeted belief-correction interventions across contexts, explore causal pathways between norms, policies, and individual action over time, and integrate these behavioral mechanisms into system dynamics and social-climate models to assess policy effectiveness.

- Survey mode and response rates: Mixed modes (telephone and face-to-face) with median response rates of 9% (telephone) and 65% (face-to-face) may introduce mode-related differences despite weighting adjustments. - Item coverage: The government action question could not be fielded in Myanmar, Saudi Arabia, and the United Arab Emirates. - Implementation errors: Minor questionnaire implementation issues occurred in Kyrgyzstan (n=4 Uzbek-language interviews missing a sentence) and Mongolia (wording led to <1% asked in Q1; corrected via approximation). Sensitivity analyses suggest results are robust. - Self-reports and general measures: Measures reflect stated willingness and general norms/policy preferences (validated correlations provided), which may differ from realized behaviors or specific policy preferences. - Cross-sectional design: Associations with GDP, temperature, and vulnerability are correlational; unobserved factors may remain despite robustness checks.