Main

Indigenous peoples, women, low-income earners, racialized minorities and other marginalized groups face the greatest risks from climate change1,2,3. Ill-considered climate policies can worsen the disproportionate burden on such groups4. Policy responses to climate change must therefore recognize the unequal distribution of causal responsibility, impacts and coping capacities across different groups and place emphasis on promoting fair and equitable outcomes5,6. Against this backdrop, climate justice has become a prominent frame in climate change messaging and advocacy.

Little is known about public understanding of climate justice or how widely the concept resonates with diverse groups. Perceptions of justice and fairness have been shown to underlie pro-environmental intentions7 and public attitudes toward various policies8. Therefore, knowledge of public opinion about climate justice can benefit efforts to promote climate action and widen support for just climate policies.

Defining climate justice

Climate justice has its roots in the anticolonial struggles of Indigenous communities of the global south9. It is also linked to the US environmental justice movement which emerged in response to the disproportionate exposure of communities of colour to harmful environmental pollution10. Popular usage of the term reflects a fusion of ecological and social justice concerns11,12, whereby disproportionate impacts on marginalized communities are highlighted and priority is accorded to protecting the rights of vulnerable groups13.

Multiple definitions of climate justice are presented in the academic literature. All reflect a consensus that climate justice encompasses, amongst other elements, notions of the distributional, procedural and recognitional (in)justices associated with climate change impacts and societal responses14,15. Distributional injustice refers to the uneven distribution of climate change vulnerability and adaptive capacity across different social groups. This is broadly linked to histories of colonization; racial, economic and political oppression13,16; and the unequal distribution of wealth and power between and within societies17. Procedural injustice captures the under-representation of frontline communities in the selection and implementation of responses to climate change. Finally, recognitional injustice describes a failure to recognize the rights and interests of marginalized groups as legitimate in climate change-related policy-making14. Misrecognized groups may be further disenfranchized by climate change adaptation and mitigation efforts18. Climate justice offers a framework for identifying and tackling the myriad ways in which the climate crisis aligns with long-standing and interconnected patterns of social injustice.

Public understanding of climate justice

In a 2023 study, roughly two-thirds of Americans (65%) had never heard of climate justice, even though more than half (53%) were supportive of the concept after it was explained to them19. In the United Kingdom, the Framing Climate Justice project found that many people recognized the unequal distribution of climate change vulnerability between high- and low-income countries but few understood how climate change relates to racial and class inequalities20.

Despite endorsing related facts such as the disproportionate burden of climate change on those who are least responsible, most participants were unable to define climate justice in a study of 6,000 young people aged 18 to 35 years in six European countries (Czechia, Germany, Italy, Romania, Spain and the United Kingdom)21. Many of these Europeans were unaware of the interconnections between climate change and issues such as gender or racial inequalities21. Similar patterns emerged in US research where there is greater recognition of the economic dimensions of climate justice than the racial dimensions among the public19,22.

Previous research offered insight into public understanding of climate justice in Europe and North America but little indication of how widely the findings generalize around the world.

Linking climate justice beliefs to action and policy support

Justice perceptions can serve as a bridge or barrier to cooperation23. For example, environmental justice beliefs and perceived fairness are linked to greater acceptance of climate policies8. Moral concerns (such as concern about negative impacts on future generations) predict greater climate policy support24. Climate activists also commonly report justice-seeking as their motivation25.

In contrast, justice considerations may prompt counterproductive responses. When players in an economic game received information about historical carbon emissions, successors of high emitters offered to pay more to mitigate climate change while successors of low emitters offered to pay less26. Climate narratives focusing on social justice polarized a UK audience along political lines; eliciting strong acceptance among politically left-leaning individuals and reactance among those who were right-leaning27. Furthermore, a US study revealed that climate justice proved too complex a concept in a climate advocacy workshop; leaving some participants unmotivated to act despite learning about ongoing climate injustices in their local area28.

Intuitively, awareness of climate (in)justices is unlikely to motivate action among groups that benefit from the underlying structural inequalities and may inhibit willingness to act among disadvantaged groups23,29. Tolerance of inequality in society and a desire to be socially dominant are generally associated with low pro-environmental motivation and a greater inclination towards unsustainable exploitation of the environment30,31. Climate justice is therefore also unlikely to resonate with people who have a dispositional preference for social inequality. Beyond mapping public understanding of climate justice, it is important to determine how climate justice beliefs relate to climate action and policy support among different groups.

Research design

Here, we present findings from a global survey of climate justice awareness and knowledge. Data were gathered in 2022 with an online survey of adults aged 18+ years (n = 5,627) in 11 countries (Australia, Brazil, Germany, India, Japan, the Netherlands, Nigeria, Philippines, United Arab Emirates, the United Kingdom and the United States). We examined how climate justice beliefs relate to climate action and policy support using a robust psychometric instrument designed to assess public endorsement of climate justice-related beliefs. See Supplementary Table 1 for country-wise demographic profiles of the study participants.

Climate justice awareness and knowledge

Across all 11 countries, two-thirds of respondents (66.2%) indicated that they had never heard of climate justice before participating in this study (Fig. 1). Furthermore, less than one-fifth of respondents (17.2%) felt that they have ‘a fair amount’ or ‘a lot’ of knowledge about climate justice (Fig. 1). Levels of climate justice awareness were highest in India, where more than half of the sample (56.5%) reported that they had heard of climate justice, and lowest in Japan (13.8%).

Fig. 1: Distribution of climate justice awareness and knowledge by country.
figure 1

Climate justice awareness and knowledge by country.

Full size image

Endorsement of climate justice beliefs

A climate justice beliefs index (CJBI) was developed to capture public endorsement of beliefs about the disproportionate impacts of climate change on women, the poor and people of colour; the need for greater representation of frontline communities in climate change-related decision-making; the need to redistribute resources from the rich to the poor; and the connection of climate change with colonization and capitalism (Table 1; see Methods and Supplementary Tables 2, 3 and 6 for details of CJBI factorial analysis and invariance testing).

Table 1 Climate justice beliefs index
Full size table

The CJBI provided insight into public endorsement of key ideas commonly associated with climate justice (Tables 1 and 2). The belief that poorer people suffer worse impacts from climate change (78% agreement) and that people from the worst-affected communities should have more of a say in decisions concerning climate change (78% agreement) were most widely endorsed. The lowest levels of endorsement were recorded for items capturing the gender and racial dimensions of climate (in)justice. Roughly equal proportions of respondents agreed (41%) or disagreed (42%) that climate change affects women worse than men around the world and just over half of all respondents (57%) agreed that Indigenous people and people of colour are worse affected. In line with scholarly perspectives on climate coloniality13, recognition of capitalism and colonialism as underpinning elements of the climate crisis (70% agreement) was prevalent across countries (Tables 1 and 2).

Table 2 Agreement with CJBI by country
Full size table

Cross-national variance in CJBI scores was low (4%). The aggregated (mean) scores across the nine CJBI statements indicate that the notions of climate justice captured by the scale are generally supported in all countries (Supplementary Table 2). The CJBI scores showed small to moderate positive correlations with climate justice awareness and knowledge in all countries except Germany and Japan. This suggests that people who are aware of climate justice also tend to endorse ideas commonly associated with the concept.

Predicting climate justice awareness, knowledge and beliefs

Multilevel linear modelling was used to assess how climate justice awareness, knowledge and beliefs relate to sociodemographic (age, gender, education and political orientation), psychological (climate change awareness and experience) and informational (access to information via television, radio, newspapers, Facebook, Twitter, YouTube, books, scientific journals, family and friends) predictors.

Age was inversely related to climate justice awareness. Respondents aged 25–34 years reported the highest frequencies of climate justice awareness, closely followed by the 18–24 years age group (Table 3 and Supplementary Table 4). However, the 18–24 years group reported the lowest self-ratings of their climate justice knowledge and showed lower endorsement of the climate justice beliefs than all but the oldest age cohorts (45–54, 55+ years). There were no significant gender differences in climate justice awareness and endorsement of climate justice beliefs. Education (attaining a college degree or higher) showed positive associations with climate justice awareness, knowledge and CJBI score. Political orientation was positively related to climate justice knowledge and negatively related to CJBI score. This means that, on average, politically right-leaning individuals provided higher self-ratings of their climate justice knowledge but indicated lower endorsement of climate justice beliefs.

Table 3 Predictors of climate justice awareness, knowledge and beliefs
Full size table

Respondents who rated themselves as being more informed about climate change were also more likely to have previously heard of climate justice, have greater climate justice knowledge and have higher CJBI scores than those who rated themselves as less informed about climate change (Table 3). However, while perceived personal experience of climate change was positively associated with CJBI scores, it showed no significant relationship with climate justice awareness and knowledge.

Climate justice awareness, knowledge and beliefs showed varying patterns of association with different climate change information sources (Table 3). People who mainly get their information from television reported lower climate justice awareness and knowledge, while those who get their information from newspapers reported higher awareness and knowledge. Facebook and Twitter use were positively associated with climate justice awareness and knowledge, while YouTube use only showed a positive relationship with climate justice knowledge and endorsement of climate justice beliefs. Accessing information from books and magazines or scientific journals and blogs was positively associated with climate justice awareness and knowledge, whereas getting climate change information from friends, family and colleagues was only positively associated with endorsement of climate justice beliefs.

Climate justice beliefs linked to actions and policy support

We also examined how climate justice beliefs relate to climate action (for example, attending protests), online activism (for example, signing online petitions to promote climate change mitigation), private-sphere pro-environmental behaviours (PEB; for example, saving energy in the household) and support for ‘push’ climate change mitigation policies (for example, increasing taxes on fossil fuels) in different countries. Descriptive statistics for the behavioural and policy support measures are provided in Supplementary Table 3.

After adjusting for the demographic covariates (age, gender and education), climate justice beliefs positively predicted engagement in climate action, online activism, PEB and support for climate change mitigation policies (Table 4). Notably, there was no significant cross-national variation in the magnitude or direction of the association between climate justice beliefs and engagement in climate action (Fig. 2). However, the relationship between climate justice beliefs and engagement in online activism was weaker in the Philippines than the overall average and stronger in Australia, Brazil and the United States. The relationship between climate justice beliefs and PEB was also weaker than average in Nigeria and the Philippines but stronger in Australia, Japan and the United States. Interestingly, climate justice beliefs showed a weaker association with climate change mitigation policy support in India, the Netherlands, Nigeria and the Philippines than average, while showing a stronger relationship in Australia, Brazil, the United Kingdom and the United States.

Table 4 Climate justice beliefs as a predictor of climate action and climate policy support
Full size table
Fig. 2: Country-wise random effects intercept and random slope estimates for climate action, online activism, PEB and push policy support regressed on climate justice beliefs, controlling for age, gender and education.
figure 2

Values represent the variance of each country’s intercept and slope from the estimated fixed effects. Error bars are country-wise random intercept and slope variance estimates ±1 s.d. Note: cjbi_gm = grand mean-centred CJBI scores. Analytical sample sizes: n = 3,649 (climate action), n = 3,581 (online activism), n = 3,636 (PEB), n = 3,649 (push climate policy support).

Full size image

Discussion

Our findings show that public awareness of climate justice as a concept is generally low around the world. India and the United Arab Emirates showed an exception where half or more of respondents reported having previous awareness of climate justice. Notably, the samples from these two countries had the highest frequencies of university-educated respondents, a factor that showed positive associations with climate justice awareness, knowledge and endorsement of climate justice beliefs. Higher education entails greater exposure to information that can foster conceptual awareness of climate justice but it is challenging if the result is that less privileged groups, those who are often most affected, are left out of the discussion.

Despite low levels of climate justice awareness, climate justice beliefs appear to be widely endorsed across countries; especially beliefs about the unequal distribution of climate change impacts, the importance of giving voice to frontline communities, the underpinning roles of colonialism and capitalism in the climate crisis and the need to redistribute resources from the rich to the poor. Furthermore, we observed positive associations between climate justice beliefs, awareness and knowledge, which suggests that climate justice beliefs among the public may be further heightened through awareness-raising campaigns.

Climate justice beliefs were robustly associated with engaging in climate action, online activism, PEB and showing support for just climate policies. The magnitude of these associations varied across countries, with the strongest associations most frequently observed in Australia, Brazil and the United States. This aligns with previous indications that awareness of historical and present climate inequalities can foster climate change mitigation actions among citizens in developed countries26. In contrast, climate justice beliefs showed weaker associations than average with online activism, PEB and mitigation policy support in Nigeria and the Philippines. We propose that climate justice beliefs may be more closely linked to individual climate action and mitigation policy support in countries with greater historical contributions to global GHG emissions and where social inequalities (for example, the disadvantaged status of racialized minorities) are more politically salient. Previous research shows that people in developing countries are less willing to make sacrifices for environmental protection as they tend to believe that wealthier (high-emitting) countries should have a greater share of the responsibility for addressing global environmental degradation32. It has been argued on this basis that invoking justice and fairness considerations could undermine environmental action in such contexts32,33. We observed varying levels of climate action, online activism, PEB and climate policy support across countries but these were often higher among people from developing countries (Supplementary Table 3). Our data suggest that people from developing countries are no less motivated to act on climate change, although climate justice beliefs may constitute a smaller part of their motivation.

It is also plausible that climate (in)justice beliefs correspond with different perspectives in different countries whereby people in historically high-emitting countries may be more likely to see themselves as a ‘perpetrator’, ‘beneficiary’ or ‘observer’ of climate injustice, while people in developing countries more commonly see themselves as ‘victims’ of climate injustice. Compared with seeing oneself as a victim of injustice, research suggests that prosocial behaviour is more likely to arise from perceiving oneself as contributing to, passively benefiting from or merely observing injustices34,35. Further research is needed to better understand the mechanisms underlying relationships between climate justice beliefs and climate action and policy support in different countries.

Stronger endorsement of climate justice beliefs was associated with climate change awareness and perceived personal experiences of climate change. Collective action by climate activist groups has raised public awareness of climate change in recent years36. Personal experiences with intensifying climate change impacts, such as extreme weather, are also changing people’s attitudes and increasing their concern about climate change37. These trends can be leveraged to widen public awareness of climate justice. For example, by highlighting social disparities in the impacts of severe weather events and drawing attention to the structural inequalities that underlie these disparities.

Our data reflect similar patterns to the politically polarized responses to social justice framing of climate change observed in previous research27. People on the right wing of the political spectrum indicated lower endorsement of climate justice beliefs than those on the left. CJBI items referring to the disproportionate impacts of climate change on women and people of colour received the lowest levels of endorsement. Views on equal rights for historically marginalized groups are ‘culture war’ issues that commonly divide political conservatives and progressives38,39. These topics may more likely prompt reactance, especially among politically right-leaning individuals. The pattern of low endorsement suggests that the connections of climate change to gender and racial injustices may not be apparent to most people.

Media sources, especially social media use and discussing climate change with family, friends and colleagues showed significant associations with endorsement of climate justice beliefs. Social media is a key channel of communication and mobilization for climate activist groups40,41. This may explain the link between climate justice beliefs and the use of Twitter or YouTube. Discussing climate change with friends and family has previously been shown to increase acceptance of climate science facts and to reinforce climate change worry42. Our data suggest that this effect also extends to climate justice-related beliefs.

A limitation of the current study is that we used samples of individuals with digital access which limits the representation of less privileged groups or people potentially at greatest risk of facing climate injustices. More inclusive data-gathering approaches are needed in future research to better represent climate justice perceptions among the most climate-vulnerable groups. We are also unable to establish causal relationships between climate justice beliefs and climate action and policy support given the cross-sectional research design. Experimental and longitudinal research approaches are needed to extend this research. Finally, climate justice issues have varying histories, manifestations and discourses across different spatial and geographical contexts43,44,45,46,47. Research should explore if, and how, varying foci of climate justice beliefs (for example, focusing on local versus global issues) shape action tendencies and policy support.

Our study provides a robust assessment of public understanding of climate justice around the world and the implications for climate action and policy support. We found that basic recognition of the social, historical and economic injustices that characterize climate change is common among the public, even if people do not consciously draw connections across these injustices using a conceptual framework of climate justice. Importantly, endorsement of climate justice beliefs is associated with engagement in climate actions and showing support for just climate policies.

Methods

Data collection

We contracted a commercial partner (Qualtrics Research Services) to recruit samples of adults aged 18+ years from the general population of Australia, Brazil, Germany, India, Japan, the Netherlands, Nigeria, the Philippines, the United Arab Emirates, the United Kingdom and the United States to complete an online survey. The countries were selected to reflect populations with a diverse range of cultures, income, climate change vulnerabilities and responsibilities for global GHG emissions. Sampling was stratified by age and gender. A target of 500 complete responses was set for each location on the basis of the available budget. The survey was fielded from 26 May to 30 June 2022.

The survey included an attention check and respondents who failed this were screened out of the study. We also screened out respondents who completed the survey in less than half the median completion time. Respondents who provided poor-quality responses (for example, irrelevant answers and straight-lining) were scrubbed and replaced by the survey company. Summaries of the country-wise effective samples and demographic breakdown are provided in Supplementary Table 1. Importantly, while online samples provide good quality data, are considered appropriate for preliminary research and are common in the academic literature48, readers should be mindful that the respondent samples recruited for this study are not fully representative of the general population of the countries from which they were drawn.

Our questionnaire was developed in English and translated into Arabic, Brazilian Portuguese, Dutch, German and Japanese by the project team working in pairs including at least one fluent bilingual speaker. The translation was completed using the translate-back translate method49. The questionnaire was hosted on the Qualtrics online platform and presented to respondents in their local language through automatic detection of the respondents’ browser settings. Respondents could also select their preferred language from a pull-down menu presented on the first page of the questionnaire.

Respondents were given information about the nature, purpose and possible risks associated with participating in the study before being asked for consent to participate. Those who did not give consent were redirected to the end of the study. Respondents were further allowed to terminate their participation at any point and at the end were provided with a debrief that included a description of the purpose of the study as well as links to websites where they could get further information about climate change and climate justice.

Outcome measures

We developed three measures to capture public attitudes about climate justice in this study: climate justice awareness, self-rated climate justice knowledge and the CJBI. In addition, we used measures aimed at capturing behavioural responses to climate change and climate policy support, particularly real-world climate actions, online climate activism, private-sphere PEB and support for radical climate change policies designed to curb high-carbon emission behaviours.

The climate justice awareness measure was a single question: ‘Before today, have you ever heard the term climate justice?’ Responses were indicated as a ‘no’ or ‘yes’, which were coded for analysis as ‘0’ or ‘1’, respectively. Self-rated climate justice knowledge was also measured with a single question: ‘How much do you know about climate justice?’ Responses were indicated on a four-point scale: ‘nothing at all’, ‘a little’, ‘a fair amount’ and ‘a lot’, which were coded as ‘0’, ‘1’, ‘2’ and ‘3’, respectively.

The CJBI comprised nine items listed below and responses were recorded on a four-point response scale (1, strongly disagree; 2, tend to disagree; 3, tend to agree; 4, strongly agree) and a ‘don’t know’ option which was coded as missing. Responses to all nine items were averaged to form an aggregate CJBI score for each respondent.

  1. (1)

    People living in poverty suffer worse effects from climate change.

  2. (2)

    Around the world, people who are least responsible for causing climate change suffer the most severe climate change impacts.

  3. (3)

    Climate change affects women worse than men around the world.

  4. (4)

    Climate change will worsen existing oppressions and inequalities (for example, the gap between rich and poor countries).

  5. (5)

    The negative impacts of climate change are worse for Indigenous people and people of colour (for example, Black, Asian and Middle Eastern) around the world.

  6. (6)

    Solving climate change requires redistributing resources from the wealthy to those who have less.

  7. (7)

    People from communities most affected by climate change should have more of a say in decisions about solutions to climate change than they currently do.

  8. (8)

    Climate change is driven by exploitative systems like capitalism.

  9. (9)

    Colonization and historical practices of forced extraction of resources from colonized territories have played a significant role in driving climate change.

These items were determined to be suitable indicators of climate justice beliefs based on a review of communications by climate change and climate justice activists around the topic50,51,52. The Framing Climate Justice project20 was a key reference for developing the CJBI.

Our climate action scale comprised six items with a yes/no response format. Respondents were asked if they had engaged in the following activities in the past year:

  1. (1)

    Attended a climate protest.

  2. (2)

    Donated money to an organization tackling climate issues.

  3. (3)

    Volunteered in an organization tackling climate issues.

  4. (4)

    Signed a petition or contacted a politician about climate change.

  5. (5)

    Attended a public lecture, seminar or workshop about climate change.

  6. (6)

    Are active members of a group that addresses climate change.

The responses were coded as yes (1) and no (0). Responses across all six items were averaged to form an aggregate climate action score for each respondent.

The online activism scale comprised four items with a five-point response format (1, almost never; 5, almost always). Selection of a ‘don’t know’ response was coded as missing. Respondents were asked to indicate how often they do the following:

  1. (1)

    Follow Facebook, links, webpages and/or Twitter accounts addressing climate change.

  2. (2)

    Express support of efforts to address climate change on social media (Facebook, Twitter, blogs).

  3. (3)

    Sign online petitions to support climate change mitigation.

  4. (4)

    Join groups online to support climate change mitigation.

Responses to these scale items were averaged to form one aggregate online activism score per respondent.

The PEB scale was adapted from ref. 53. The scale comprised six items with a five-point response format (1, almost never; 5, almost always). Selection of a ‘don’t know’ response was coded as missing. Respondents were asked to indicate how often they do the following:

  1. (1)

    Cycle or walk instead of driving or being driven in a car.

  2. (2)

    Buy second-hand or used items instead of buying new things.

  3. (3)

    Try to influence family members and friends to act in a climate-friendly way.

  4. (4)

    Save energy in the household.

  5. (5)

    Take public transportation instead of the car.

  6. (6)

    Avoid food waste.

Responses to these scale items were averaged to form one aggregate PEB score per respondent.

The climate change mitigation (push) policy support scale was designed to capture support for radical policies intended to reduce demand for, or push consumers away from, high-carbon behaviours and practices. Respondents were asked about the extent to which they support or oppose the following policies:

  1. (1)

    Increasing taxes on fossil fuels (for example, coal, oil, diesel, petrol and gas).

  2. (2)

    A ban on production of new cars that are powered by an internal combustion engine or fossil fuels (for example, petrol, diesel and gas) by 2030.

  3. (3)

    Increasing taxes on carbon-intensive foods such as meat and dairy (milk products).

  4. (4)

    Additional charges for people who fly more than twice a year (a ‘frequent flyer’ levy).

Responses to these items were recorded on a four-point scale (strongly oppose, 1; tend to oppose, 2; tend to support, 3; strongly support, 4). Responses to these scale items were averaged to form one aggregate push policy support score per respondent.

Predictor measures and covariates

We identified potential predictors of climate justice awareness, knowledge and beliefs on the basis of three domains shown to be important determinants of climate perceptions in previous research: demographic factors (age, gender and education), identity and ideology (political orientation), knowledge (self-rated climate change information and information sources) and experience54,55. The demographic factors were also included as covariates when assessing how climate justice beliefs relate to climate action and policy support.

Respondents were asked to report their age in years. They were also asked to indicate if they identify as: ‘woman’, ‘man’, ‘other (specify)’ or ‘prefer not to say’. Responses were recoded for analysis whereby those who identify as ‘woman’ were coded as ‘1’ and all other responses were coded as ‘0’. We also asked respondents to indicate their highest level of educational attainment by selecting one of six categories (no formal education, primary school, secondary school or vocational equivalent; community college, trade school or vocational equivalent; university degree or higher or other). The ‘university degree or higher’ response was coded as ‘1’ and other responses were coded as ‘0’.

Political orientation was measured with a single question: ‘In politics, people often talk about the ‘left wing’ and the ‘right wing’, Below is a scale where 0 represents those who are on the far left while 10 represents those who are on the far right. Where would you place yourself on such a scale?’56

Respondents’ self-rated level of information about climate change was measured with a question: ‘How informed would you say you are about climate change?’ Response options were ‘not at all informed’, ‘a little informed’, ‘well informed’ and ‘extremely well informed’. These were coded as ‘1’, ‘2’, ‘3’ and ‘4’, respectively.

Perceived personal experience of climate change was measured with a question: ‘Have you experienced any direct or indirect effects of climate change in your lifetime?’ Response options were ‘no’ and ‘yes’ which were coded as ‘0’ and ‘1’, respectively.

Finally, respondents were asked to report the frequency at which they access climate change-related information from a range of sources encompassing both traditional (television, radio, newspapers, books and magazines) and social media (Facebook, Twitter and YouTube), as well as scientific journals or blogs and family and friends. Responses were recorded on a nine-point scale ranging from ‘never’ to ‘more than ten times per day’.

Psychometric analysis

Reliability

Descriptive statistics and indices of reliability (Cronbach’s alpha) for the CJBI are presented in Table 3. The scale showed high reliability across all countries. It was positively correlated with climate justice awareness and self-reported climate justice knowledge in most countries except Germany and Japan. The other multi-item scales—climate action, online activism, PEB and push policy support also showed high levels of reliability (Supplementary Table 2) as well as significant positive correlations with CJBI (Supplementary Table 4).

Factor analysis

CJBI

Using a ‘hold-out’ sample of 549 cases randomly selected from the full cross-national dataset, we conducted an exploratory factor analysis with principal axis factoring to assess the factor structure of the CJBI. This hold-out sample corresponds to ~10% of the total sample. The results indicate that the sampling was adequate for factor analysis of the CJBI (Kaiser-Meyer-Olkin (KMO) = 0.93, χ2(36) = 1,386.03, P < 0.001). A single latent factor was found to underlie the nine items of the CJBI, which explained 49.09% of the variance. All scale items showed high loadings (>0.60) on to this latent factor.

Subsequently, we conducted multigroup confirmatory factor analysis (CFA) to determine if the single latent factor structure is supported in all 11 countries represented in the study. The results of the multigroup CFA showed a good fit, which indicates that the model has equivalent form or configural invariance across all countries (χ2(297) = 177.09, P = 1.000, CFI = 0.95, root mean square error of approximation (RMSEA) = 0.054, standardised root mean square residual (SRMR) = 0.035). The scale also showed metric invariance, which suggests that the scale items are interpreted in a similar fashion in all countries (see Supplementary Table 5 for metric invariance fit indices).

Climate action scale

We also used the hold-out sample (n = 549) to conduct an initial exploratory factor analysis of the climate action scale. The results indicated that the sampling was adequate for factor analysis of this scale (KMO = 0.85, χ2(15) = 1,023.14, P < 0.001). A single latent factor was observed and this explained 43.9% of the variance across the six items comprising the scale. All items showed moderate to strong loadings on this latent factor (>0.50).

Next, we conducted a multigroup CFA to determine if the single latent factor structure was supported in all 11 countries. The results of the multigroup CFA indicate that the model has configural invariance across all countries (χ2(99) = 146.17, P < 0.001, CFI = 0.97, RMSEA = 0.051, SRMR = 0.034). The scale also showed metric invariance across countries (Supplementary Table 5).

Online activism scale

The results of an exploratory factor analysis indicated adequate sampling (KMO = 0.84, χ2(6) = 1,209.52, P < 0.001) and a single latent factor which explained 71.8% of the variance across the items comprising the scale. All items loaded strongly onto this latent factor (>0.70).

We also conducted a multigroup CFA to determine if the single latent factor structure was supported across the 11 countries. The results showed an acceptable fit with the data (χ2(22) = 13.87, P = 0.906, CFI = 0.99, RMSEA = 0, SRMR = 0.014). Therefore, we concluded that the scale has configural invariance. The scale also showed metric invariance (Supplementary Table 5).

PEB scale

Exploratory factor analysis with the hold-out sample revealed that the sampling level was slightly lower than the conventional threshold (KMO ≥ 0.80) of sampling adequacy (KMO = 0.77, χ2(15) = 762.78, P < 0.001). Further, two latent factors were identified within the scale, which together explained 49.8% of the variance across all six items. The items measuring ‘use of public transportation instead of car’, ‘cycle or walk instead of driving or being driven’, ‘try to influence friends and family to act in climate-friendly way’ and ‘buy second-hand items instead of new ones’ loaded on to the first factor—PEB1. The items measuring ‘avoid food waste’ and ‘save energy in the household’ loaded onto the second factor—PEB2.

Multigroup CFA on the full dataset revealed that the two-factor structure of the scale had a good fit with the data and was supported in all 11 countries (χ2(88) = 221.50, P < 0.001, CFI = 0.98, RMSEA = 0.059, SRMR = 0.044). The scale also showed metric invariance (Supplementary Table 4).

Climate change mitigation (push) policy support

Exploratory analysis with our hold-out sample showed adequate sampling (KMO = 0.80, χ2(6) = 743.14, P < 0.001) and a single latent factor that explained 71.8% of the variance across the scale items. All items loaded strongly onto this latent factor (>0.65).

Multigroup factor analysis revealed that the single factor structure was supported across all 11 countries (χ2(22) = 9.95, P = 0.987, CFI = 1.00, RMSEA = 0, SRMR = 0.012). The scale also showed metric invariance (Supplementary Table 5).

Analytic strategy

Preparation of the data for analysis was done using SPSS v.27. We investigated predictors of climate justice awareness, self-rated knowledge and beliefs, as well as relationships between climate change justice beliefs and the behavioural and policy support outcomes with hierarchical linear models using the R package lme4 for mixed effects models57.

In an initial round of analysis, we estimated the intraclass coefficients (ICC) for all outcome variables to determine the magnitude of clustering effects present in the data. We found that country explained 8% of the variance in climate justice awareness, 14% of the variance of the variance in self-rated climate justice knowledge and 4% of the variance in climate justice beliefs. We also found that country explained 18% of the variance in climate action, 22% of the variance in online climate activism, 16% of variance in PEB and 6% of the variance in push policy support. Although, the ICC estimates vary across the different variables, the data show moderate to strong clustering effects on average. Therefore, the use of a multilevel modelling approach is justified.

For the models predicting climate justice, self-rated knowledge and beliefs, we estimated the linear fixed effects of the predictors and random intercepts for the grouping variable—country. For models assessing how climate justice beliefs relate to the behavioural and policy support outcomes, we were also interested in determining if the effects of climate justice beliefs vary across countries. Therefore, we added a random slope for CJBI, alongside fixed effects for the demographic covariates and random intercepts for country. We compared this model with a simpler model without random slopes for CJBI using a likelihood ratio test to determine if the slopes for CJBI vary significantly across countries for each outcome variable. These comparisons showed that adding a random slope for CJBI to the model predicting climate action did not differ significantly from the simpler model containing only a fixed effect of CJBI (Χ2(2) = 0.72, P = 0.698). In other words, the relationship between climate justice beliefs and climate action does not vary significantly across countries. In contrast, the random slopes of CJBI were significant in the models predicting online activism (Χ2(2) = 21.01, P < 0.001), PEB (Χ2(2) = 20.43, P < 0.001) and push policy support (Χ2(2) = 64.11, P < 0.001), suggesting that the relationship between climate justice beliefs and these outcome variables varies significantly across countries.

Data transformation and missing data

For the multilevel models, all predictor variables were grand mean-centred before analysis. ‘Don’t know’ responses were treated as missing data and listwise deletion was used to address missing responses across all of the statistical analyses.

Ethics statement

Ethical approval for this study was granted by the University of Nottingham School of Psychology Research Ethics Committee (ref. F1344). All participants provided informed consent before completing the study.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.