Introduction

Climate change poses unprecedented challenges for humanity, with future generations bearing the brunt of its impacts. Young people are expected to play a central role in mitigating, adapting to, and addressing the consequences of climate change. Climate change education (CCE), which equips individuals with the knowledge, attitudes, and skills needed to address climate challenges1, is essential for fostering informed decision-making and transformative action2. Learner-centred approaches in CCE, which emphasise tailoring educational content to students’ pre-existing knowledge and experiences3, require a nuanced understanding of young people’s conceptions of climate change—encompassing their knowledge, attitudes, and behaviours4. However, the extent to which these conceptions have been explored, particularly in developing countries like in Southeast Asian countries, remains limited.

Research on CCE has predominantly focused on Western and Global North contexts, often overlooking socio-cultural factors that shape climate change understanding in the Eastern and developing contexts. As a socio-scientific issue, climate change is deeply intertwined with social, economic, cultural, and political factors, making context-specific educational approaches crucial5,6. Tailoring CCE to a specific region requires an understanding of local knowledge systems, cultural values, and socio-economic realities. In Indonesia, these factors are particularly significant due to the country’s unique vulnerability to climate change and its critical role in global climate action7.

Indonesia, an archipelagic nation with over 17,500 islands and 80,000 kilometres of coastline, is among the countries most vulnerable to climate change due to its varied geography, topography, and climate. Frequent climate-related events such as floods, droughts, and rising sea levels already affect millions of Indonesians, while future risks threaten the livelihoods and well-being of its population7. Despite these vulnerabilities, the country’s preparedness remains at a lower-middle level8. As the fourth most populous nation and one of the largest greenhouse gas emitters, Indonesia has committed to significant climate action, including a 43% emission reduction target with international support and achieving net-zero emissions by 2060. Achieving these ambitious goals requires greater public awareness and engagement, particularly among adolescents, who will inherit the responsibility of sustaining these efforts.

Many studies have examined adolescents’ conceptions of climate change, focusing on awareness, knowledge, attitudes, and behaviour4,9,10,11,12,13,14,15. However, the literature is dominated by studies from Western and/or the Global North contexts and tend to overlook the socio-cultural contexts that shape the understanding of climate change. In Indonesian context, studies16,17,18 have predominantly focused on the cognitive domain—knowledge and awareness—while often neglecting the affective and behavioural dimensions that are critical to fostering meaningful climate action.

This study aims to fill these gaps by investigating the conceptions of climate change among Indonesian upper-secondary students, incorporating cognitive (perceived knowledge, factual knowledge, and belief), affective (attitude, concern, and hope), and behavioural dimensions (willingness to act and self-efficacy), and situating the findings within the broader socio-cultural context of Indonesia. This study is guided by three key questions: 1) What are Indonesian upper-secondary students’ conceptions of climate change in relation to its causes, impacts, and solutions? 2) How can Indonesian upper-secondary students be categorised into distinct groups based on their conceptions? 3) How do background factors, such as demographic characteristics, sociocultural influences, and personal values, shape their conceptions?

This paper highlights the importance of understanding conceptions of climate change among adolescents as a foundational step towards effective, learner-centred CCE. By focusing on Indonesia—a highly vulnerable yet underrepresented context in CCE research—this study addresses critical gaps in the literature, offering insights that are both contextually grounded and globally significant. The use of recent, primary data ensures the findings are timely and reflective of the lived experiences of Indonesian adolescents. Furthermore, the study’s integration of cognitive, affective, and behavioural dimensions provides a multidimensional understanding of young people’s conceptions, advancing theoretical frameworks in the field. These findings not only inform culturally and contextually tailored educational strategies but also contribute to the development of scalable approaches adaptable to similar socio-economic and cultural contexts.

Results

Demographic overview

Table 1 summarises the demographic characteristics of respondents, showing a relatively balanced distribution across grades. However, the sample is slightly skewed towards certain groups, with Muslims comprising approximately 80% of respondents, Javanese around 52%, and females about 60%. This diverse yet representative sample offers valuable insights into the climate change conceptions of Indonesian students from various socio-demographic backgrounds.

Table 1 Demographic information of respondents
Full size table

Cognitive dimension

The cognitive dimension encompasses knowledge, perceived knowledge, and belief. Respondents were questioned on their perceived knowledge of climate change and its causes, impacts, and solutions. Approximately 75% exhibited confidence in their knowledge to some extent, although more neutral and negative trends were observed in perceived knowledge about solutions. To evaluate actual knowledge, a factual knowledge test was conducted, divided into system knowledge (general understanding of climate change), action knowledge (knowledge of possible actions), and effectiveness knowledge (understanding the relative impacts of behaviours). The results, depicted in Fig. 1, revealed an average score of 43.32 (SD = 7.17) out of a maximum of 100 points. Particularly, system knowledge scored the highest, with an average of 62.29, compared to 38.62 for action knowledge and 45.86 for effectiveness knowledge.

Fig. 1
figure 1

Factual knowledge test results.

Full size image

Furthermore, a linear relationship was identified between perceived knowledge and factual knowledge scores. Respondents with higher perceived knowledge performed better on the test, achieving an average score of 43.92, compared to 39.86 for those with lower perceived knowledge. This finding highlights the connection between self-assessed confidence and actual understanding.

Several misconceptions were identified among respondents. In the system knowledge section, many believed that climate change is solely defined as the rising temperature of the Earth’s surface (49.4% incorrect, 12.7% ‘don’t know’) and that the thinning of the ozone layer is a cause of climate change (82.8% incorrect, 10.7% ‘don’t know’). In the effectiveness knowledge section, misconceptions included the relative energy use of producing virgin paper versus recycled paper (45.9% incorrect, 26.9% ‘don’t know’) and greenhouse gas emissions from short-haul versus long-haul flights (54.4% incorrect, 34.2% ‘don’t know’).

Respondents were also asked to rank seven behaviours and technologies based on the amount of greenhouse gas emissions they can avoid. As shown in Table 2, behaviours such as ‘wasting less food’ and ‘family planning’, which should rank at the top, were instead placed at the bottom. Conversely, less impactful actions like ‘household recycling’, ‘solar energy’, and ‘restoring tropical forests’ were prioritised.

Table 2 Action knowledge task answer
Full size table

Figure. 2 illustrates respondents’ sources of knowledge, the most common were the internet and social media (N = 4024), followed by school (N = 3957) and television (N = 1826). Additional sources included personal experiences, games, seminars, and environment-themed competitions. When asked about beliefs in anthropogenic climate change, 88% of respondents acknowledged its occurrence and attribution to human activity. However, 66.2% still expressed some belief that climate change is a natural phenomenon.

Fig. 2
figure 2

Climate change knowledge sources.

Full size image

Affective dimension

The affective dimension, or socio-emotional dimension, involved attitudes towards climate change, concerns about its impacts, and hopes for resolutions. Overall, 81.6% displayed positive attitudes towards climate change, particularly agreeing with statements like “people should care more about climate change” (92%) and “climate change is a threat to the world” (84.7%). However, the statement “it is annoying to see people do nothing for the climate change issues” received the lowest positive response (74.7%).

Concerns about climate change and its impacts were expressed by 87.5% of respondents. Slightly higher positive responses were observed for statements regarding the impacts on themselves and their families compared to broader societal impacts. Regarding hopes for climate change resolutions, only 76.7% expressed optimism, while 20.1% were neutral and 3.2% pessimistic. Notably, merely 64% expressed hope in politicians’ commitment to addressing climate change, reflecting a higher degree of hopelessness.

Behavioural dimension

The behavioural dimension consisted of willingness to act and self-efficacy—confidence in one’s ability to act. Compared to other dimensions, this aspect exhibited the lowest levels of positive responses. Actions such as walking or biking to nearby destinations, conserving water and electricity, consuming less, reducing waste, and learning about climate change were relatively popular. However, less enthusiasm was observed for using public transport (72.2%), joining climate action campaigns and events (69.6%), and advocating for governmental climate action (63.4%).

Self-efficacy positive responses were the lowest among all variables, with statements like “I am confident in my ability to take climate action” and “I am confident in my ability to engage others in climate action” receiving positive responses from solely 64.4% and 67.4%, respectively.

Trends by background variables

Multiple linear regression analysis was conducted to assess the effect of background variables on overall conceptions of climate change among respondents. The overall conceptions score was derived by calculating the average scores of the continuous variables—perceived knowledge, belief, attitude, concern, hope, self-efficacy, and willingness to act—all measured on a 7-point Likert scale. The factual knowledge score, originally measured on a 0–100 scale, was rescaled to a 7-point scale to ensure comparability before aggregation. Demographic factors included age, grade, gender, location, ethnicity, religion, personal values, past climate change experience, and school type (private or public). The analysis (F(11, 4489) = 219.658, p < 0.001, R2 = 0.350) exhibited statistically significant impacts of grade, location, ethnicity, religion, and personal values.

Students in Grade 11 (B = 0.644, p < 0.001) and Grade 12 (B = 0.825, p < 0.001) demonstrated higher levels of conceptions compared to Grade 10 students. Location also played a role, with students in Balikpapan (B = −1.401, p < 0.001) and Malang (B = −0.598, p = 0.001) exhibiting lower levels of conceptions than their peers in Jakarta. Ethnicity showed a significant effect, as Sulawesian respondents scored lower (B = −0.735, p < 0.001) compared to Javanese students. Religion was another factor, with Hindu students (B = 2.022, p = 0.039) showing higher levels of conception than Muslim students. Furthermore, students who had experienced the consequences of climate change exhibited higher conceptions (B = 1.845, p < 0.001). Finally, personal values significantly influenced conceptions, with higher levels of biospheric (B = 1.726, p < 0.001), altruistic (B = 1.094, p < 0.001), egoistic (B = 0.909, p < 0.001), and hedonic values (B = 0.223, p = .004) associated with greater understanding of climate change.

Typologies in climate change conceptions

Cluster analysis was conducted to classify respondents into distinct groups based on their conceptions of climate change, measured across cognitive, affective, and behavioural dimensions (maximum possible mean score = 21). As illustrated in Fig. 3, three clusters emerged from the analysis.

Fig. 3
figure 3

Parallel plot of clusters based on levels of cognitive, affective, and behavioural conceptions.

Full size image

The first cluster, labelled the Uncertain, represents approximately 26% of the sample (N = 1171). This group is characterised by relatively lower levels of conceptions, with mean scores of 12.66 (cognitive), 14.59 (affective), and 13.40 (behavioural). The second cluster, the Learners, encompasses 46.7% of the respondents (N = 2100). These students exhibit moderate levels of conceptions, with mean scores of 14.30 (cognitive), 17.64 (affective), and 16.72 (behavioural). Finally, the third cluster, the Advocates, comprises 27.3% of the total sample (N = 1230). This group demonstrates relatively high levels of conceptions, with mean scores of 15.52 (cognitive), 19.73 (affective), and 19.77 (behavioural).

Discussion

This study aimed to investigate the conceptions of climate change among Indonesian upper-secondary students, focusing on cognitive, affective, and behavioural dimensions, the influence of background variables, and clustering based on these conceptions.

Indonesian upper-secondary students demonstrated lower levels of cognitive conceptions compared to affective and behavioural dimensions. Despite expressing confidence in their knowledge, many students displayed limited and fragmented knowledge, echoing findings by another Indonesian study17. This was especially evident in their understanding of solutions to climate change. Confusion about the effectiveness of mitigation actions was common, with greater appreciation for familiar solutions such as household recycling, solar energy, and reforestation. These preferences likely stem from school environmental initiatives emphasising cleanliness and recycling, flora and fauna protection, and energy conservation19,20. This finding aligns with global trends, where familiarity often overshadow accurate comprehension of impactful solutions12,21.

Behavioural tendencies mirrored these patterns. Students favoured actions that were familiar, convenient, and personally relevant, such as conserving energy and water, walking or biking, and reducing waste. Conversely, less popular actions included using public transport, participating in climate action campaigns, and advocating for policy change. Similar trends in broader surveys of the Indonesian public reveal low engagement in social or political climate actions18.

Misconceptions, such as attributing climate change to the thinning ozone layer, were also prevalent. These persistent misconceptions are common among adults and young people worldwide, including in Japan12, Singapore13, Australia22, and America14,23. Echoing findings by Kundariati et al.16, most students recognised climate change as a real phenomenon predominantly driven by human activity. However, many also expressed confusions about its natural causes, reflecting a conflicted understanding of its causes. This emphasises the need for Transformative Learning Theory’s ‘disorienting dilemmas’ to challenge students’ existing worldviews and foster critical reflection24.

Despite these cognitive gaps, most students expressed significant concern about the consequences of climate change, particularly its impact on themselves and their families. Approximately 77% of students reported feeling hopeful, with higher levels of trust in their personal and collective abilities to address the issue. However, distrust in political leaders contributed to feelings of hopelessness. While concern is a key driver of pro-environmental behaviours, excessive concern can negatively affect mental health and lead to climate anxiety25. To prevent disengagement and inaction caused by such anxiety, it is essential to balance concern (negative emotions) with hope (positive emotions)26,27,28. CCE must present the reality of climate change as a ‘wicked problem’ while simultaneously fostering positive emotions such as constructive hope, which can inspire agency and resilience29.

Self-efficacy, a critical factor for fostering behavioural change, was among the lowest positive responses. Similar trends have been reported in Australia22. Although 78% of students expressed willingness to act, merely 71% felt confident in their ability to take effective action. This disparity suggests that low self-efficacy may hinder students’ capacity to translate intentions into meaningful behaviours. CCE can increase this self-efficacy by incorporating problem-based learning strategies that explore climate change solutions in the community22,29,30.

The multiple linear regression analysis revealed that past experience, grade level, location, ethnicity, and religion significantly influenced students’ conceptions of climate change. Students with prior exposure to climate-related events, such as floods or droughts, exhibited higher levels of understanding. Higher-grade students outperformed their lower-grade peers, likely reflecting increased exposure to environmental topics within the curriculum. Students in Jakarta exhibited a greater understanding than those in smaller cities, highlighting disparities in quality of education and access to information and resources31. In contrast, gender showed no significant effects, indicating that climate change conceptions may not be strongly mediated by gender differences.

The observed differences in climate change conceptions among students of different ethnicities and religions highlight the potential influence of socio-cultural factors28,32. For instance, the higher levels of conceptions among students from Hindu background may be attributed to the ecological values deeply embedded in their cultural and religious traditions33. Meanwhile, the relatively lower scores of Sulawesian students could stem from socio-economic or educational disparities that affect access to high-quality resources and information31. These findings suggest that ethnic and religious identities play a nuanced role in shaping young people’s understanding of climate issues. CCE needs to incorporate socio-cultural contexts of the learner to enhance personal relevance32. Further research should explore these dynamics through in-depth qualitative methods to uncover how cultural and religious domains interact with climate change conceptions.

Among the value dimensions, biospheric and altruistic values exhibited the strongest associations, suggesting that individuals who prioritise environmental preservation and the welfare of others are more likely to possess a greater understanding of climate change. This aligns with prior studies highlighting the role of pro-social and pro-environmental orientations in shaping environmental cognition34,35. Interestingly, egoistic and hedonic values also showed significant, although weaker, positive relationships. This implies that even self-centred or pleasure-driven values can play a role in shaping climate change conceptions, possibly because climate change impacts are perceived as threats to personal well-being or because climate action aligns with personal benefits, such as cost savings or improved quality of life.

Cluster analysis identified three distinct groups: Learners, Advocates, and the Uncertain. Nearly half of the students belonged to the Learners group, characterised by moderate levels of cognitive, affective, and behavioural conceptions. More than 27% of students were in the Advocates group, reflecting a hopeful trend that Indonesia’s younger generation is increasingly aware, emotionally engaged, and proactive. This positive trajectory can be nurtured through the combined efforts of schools, families, media, and communities36. Nevertheless, special attention needs to be paid to the bottom 26%, who require targeted interventions to improve their overall conceptions.

This classification highlights the diverse educational needs of different groups, accentuating the learner-centred approach of CCE3. While all groups would benefit from comprehensive interventions addressing scientific knowledge and the effectiveness of climate solutions, tailored approaches are essential. For instance, the Uncertain group, with lower interest levels, might benefit from creative, art-based learning methods37. Activities such as writing poems, crafting stories, or creating artwork about climate change could spark initial interest and lead to deeper engagement through artistic approaches. For the Advocates, more advanced approaches, such as problem-based, project-based or inquiry-based learning, are better aligned29,30, empowering them as change agents to explore and implement pathways for climate mitigation and adaptation.

Tailoring context-specific educational initiatives is critical for enhancing public understanding and empowerment6,32. These initiatives should integrate cognitive, affective, and behavioural dimensions to provide a holistic understanding of climate change, address misconceptions, and promote impactful solutions beyond superficial actions. Transformative learning can be applied through real-world problem-based activities and reflective discussions that encourage students to challenge pre-existing assumptions and explore effective climate solutions24. Additionally, region-specific curricula should reflect local socio-economic, cultural, and environmental contexts to ensure relevance and effectiveness. For example, incorporating indigenous knowledge systems into learning could contextualise global climate concepts.

This study offers insights into the conceptions of climate change among upper-secondary students in diverse regions of Indonesia, an underrepresented context in CCE research. The integration of cognitive, affective, and behavioural dimensions provides a comprehensive framework for analysis. However, the reliance on self-reported data may introduce biases, as responses may not always align with actual behaviours. Future studies could address this limitation by triangulating self-reported data with behavioural observations, phenomenological approaches, or in-depth interviews to provide a more holistic understanding. Additionally, the exclusion of certain rural and underserved regions may limit the depth of insights, particularly in capturing perspectives from vulnerable populations who are disproportionately affected by the consequences of climate change. Including these groups in future research would not only enhance representativeness but also bring critical insights into how socio-economic and geographical disparities shape climate change conceptions.

Methods

Data collection

The study employed quantitative approaches using questionnaire surveys targeting upper-secondary schools in four municipalities in Indonesia: Jakarta, Surabaya, Malang, and Balikpapan. These municipalities were selected to provide diverse geographical and socio-economic contexts within the country. Collaboration with the Indonesia’s Ministry of Education was instrumental in facilitating the dissemination of the online survey. With the Ministry’s assistance, schools in each municipality were invited to participate in the study between August 2024 to September 2024. Efforts were made to include schools of varying characteristics to ensure a broad representation of upper-secondary students. Schools were selected based on type (public and private), reputation (highly ranked and lower-ranked schools in terms of academic performance), location (urban and suburban), religious affiliation (secular and religious schools), and the presence of environmental-focused programmes. In total, 35 schools participated in the study, encompassing a wide spectrum of institutional profiles and academic contexts.

Students were invited to complete the survey during school hours under the supervision of their teachers or school administrators. Participation was voluntary and non-incentivised, and informed consent was obtained from students prior to data collection. Their confidentiality and anonymity were ensured throughout the research process. Participants were also given the option to withdraw from the study at any time without any repercussions. No interventions were introduced, as the study aimed to explore students’ existing conceptions of climate change in their natural learning environments. The final and valid sample consisted of 4,501 students, providing a substantial dataset for analysing adolescents’ conceptions of climate change.

Measures

This study aimed to explore students’ conceptions of climate change incorporating cognitive, affective, and behavioural dimensions. This construct is built upon the three learning dimensions of Education for Sustainable Development3 and Hilgard’s trilogy of mind38. Moreover, it accounts for background variables, including demographic and socio-cultural factors.

Cognitive dimension

The cognitive dimension focuses on understanding climate change and includes three key constructs: perceived knowledge, factual knowledge, and belief.

  • Perceived knowledge refers to an individual’s belief that they are knowledgeable about climate change. However, perceived knowledge does not necessarily equate to factual knowledge. Studies have shown that individuals with higher perceived knowledge often score lower on actual knowledge assessments39, highlighting the need for assessing both. Four items were included and measured on a 7-point Likert scale (1 = strongly disagree, 7 = strongly agree). Internal consistency was high (Cronbach’s α = 0.90).

    1. 1.

      I know about the phenomenon of climate change.

    2. 2.

      I know about the causes of climate change.

    3. 3.

      I know about the impacts of climate change.

    4. 4.

      I know about the solutions to climate change.

  • Knowledge refers to the scientific understanding of climate change—its causes, impacts, and solutions. Factual knowledge has been proved to directly and indirectly affect a person’s willingness to act12,40. It can be categorised into three types: system knowledge, which includes a general understanding of climate change, its causes, and effects; action knowledge, referring to knowledge of possible courses of action; and effectiveness knowledge, which involves understanding the relative impacts of different behaviours41,42.

    System knowledge: Six questions were included with true/false/don’t know options. Correct responses were counted as 1 point while incorrect and “don’t know” responses were counted as 0 point.

    1. 1.

      Most of the current climate change is due to greenhouse gases generated by human activity.

    2. 2.

      Climate change is only defined as the rising of temperature of the earth’s surface.

    3. 3.

      Climate change is a result of the ozone layer becoming thinner.

    4. 4.

      Rise in sea level and drought are some of the consequences of climate change.

    5. 5.

      Because of climate change, the water in seas and oceans will expand.

    6. 6.

      Because of climate change, certain plants and animals may become extinct.

    Action knowledge: Respondents were asked to rank seven options of climate change mitigative strategies in order of how much greenhouse gases can be avoided. Scores were weighted according to their ranks.

    1. 1.

      Household recycling

    2. 2.

      Plant-based diet

    3. 3.

      Family planning

    4. 4.

      Wind energy

    5. 5.

      Solar energy

    6. 6.

      Wasting less food

    7. 7.

      Restoring tropical forests

    Effectiveness knowledge: Six questions were included with true/false/don’t know options. Correct responses were counted as 1 point while incorrect and “don’t know” responses were counted as 0 point.

    1. 1.

      Consuming domestic beef uses less energy than consuming imported beef.

    2. 2.

      It takes the same amount of energy to produce recycled paper as it takes to produce conventional paper.

    3. 3.

      Energy saving light bulbs consume less energy than conventional light bulbs with the same illuminating power.

    4. 4.

      A car’s average CO2 emission per person and kilometre exceeds that of a train many times over.

    5. 5.

      The production of 1 kg of beef produces more greenhouse gases than the production of 1 kg of wheat.

    6. 6.

      On short-haul flights (e.g., within Indonesia) the average CO2 emission per person and kilometre is lower than on long-haul flights (e.g., Indonesia to America).

    7. 7.

      A diesel-engine vehicle causes more CO2 per person and kilometre than a comparable petrol-engine vehicle.

  • Knowledge source: Respondents were asked to identify their sources of information about climate change through the following question: “What sources did you get information about climate change? (Please select all that apply)”. The response options included: School, internet and social media, television, books or magazines, family, friends, and other (please specify).

  • Belief pertains to whether individuals accept the occurrence of anthropogenic climate change. Disbelief in human-caused climate change has been shown to negatively influence a person’s willingness to engage in climate-related actions6,43. Four items were included and measured on a 7-point Likert scale (1 = strongly disagree, 7 = strongly agree). Internal consistency was low (Cronbach’s α = 0.46), primarily due to the divergent responses to item 4 that was reverse coded (e.g., 1 = strongly disagree was flipped to 7 = strongly disagree). However, I chose to retain item 4, as it highlights an important aspect of respondents’ beliefs regarding the causes of climate change.

    1. 1.

      I believe climate change is happening.

    2. 2.

      I am certain climate change is happening.

    3. 3.

      I believe climate change is caused by human activity.

    4. 4.

      I believe climate change is caused by natural processes.

Affective dimension

The affective dimension refers to the emotions and feelings associated with climate change, which are among the strongest predictors of decisions and behaviours related to it5,12,25,40,44. In this framework, the affective dimension involves attitude, concern, and hope.

  • Attitude towards climate change reflects an individual’s overall evaluation of the phenomenon whether positively or negatively4,45,46. Attitudes can influence one’s behavioural intentions and willingness to engage in climate-related actions, making this dimension critical in understanding climate change conceptions. Seven items were included and measured a 7-point Likert scale (1 = strongly disagree, 7 = strongly agree). Internal consistency was high (Cronbach’s α = 0.88).

    1. 1.

      People should care more about climate change.

    2. 2.

      Climate change should be given top priority.

    3. 3.

      It is annoying to see people do nothing for the climate change issues.

    4. 4.

      The seriousness of climate change has been depreciated.

    5. 5.

      Climate change is a threat to the world.

    6. 6.

      It is my responsibility to act in a climate-friendly manner.

    7. 7.

      I can contribute to reducing the degree of climate change.

  • Concern refers to the perceived risks or threats posed by climate change. In other words, it reflects the degree of worry about its potential impacts on one’s life, family, and humanity as a whole4,12,36. Concern has been identified as one of the most significant drivers of individual support for climate change policies and engagement in pro-environmental behaviours25,44,47. Five items were included and measured a 7-point Likert scale (1 = strongly disagree, 7 = strongly agree). Internal consistency was high (Cronbach’s α = 0.94).

    1. 1.

      I am worried about climate change.

    2. 2.

      I am worried about climate change’s impacts on your life.

    3. 3.

      I am worried about climate change’s impacts on the life of your family.

    4. 4.

      I am worried about climate change’s impacts on the life of people in Indonesia.

    5. 5.

      I am worried about climate change’s impacts on the life of all humankind.

  • Hope relates to the degree of optimism or pessimism about the potential resolution of climate change26. Positive emotions, such as hope, have been found to be positively correlated with the formation of pro-environmental behaviour12,45. Hope is categorised into two categories: hope based on denial and constructive hope. Unrealistic optimism or hope based on denial has been shown to negatively affect behaviour. In contrast, constructive hope has a positive influence on pro-environmental behaviour. Constructive hope is characterised by three themes: positive reappraisal, trust in others, and trust in oneself26. Adapted her study, six items were included and measured a 7-point Likert scale (1 = not at all, 7 = very well). Internal consistency was high (Cronbach’s α = 0.82).

  1. 1.

    I feel hope because humanity has confronted complex and seemingly hopeless societal problems before and has been able to solve them eventually (positive reappraisal).

  2. 2.

    I feel hope because the awareness about this problem has increased considerably during recent years (positive reappraisal).

  3. 3.

    I feel hope because I believe that research and technical solutions will contribute to the improvement of the climate change problem (trust-others).

  4. 4.

    I feel hope because politicians in more and more countries take climate change seriously’ (trust-others).

  5. 5.

    I feel hope because we as individuals can change our behaviour; together we can influence climate change in a positive direction (trust-self).

  6. 6.

    I feel hope because I know that there are several things that I can do to contribute to the improvement of the climate change problem (trust-self).

Behavioural dimension

The behavioural dimension focuses on actions taken to mitigate or adapt to climate change, as well as the confidence to take such actions, which is referred to as self-efficacy.

  • Willingness to act is often used as a proxy for actual behaviour, as assessing real-world actions can be challenging in practice. Willingness to act has strong predictive value for future behaviours40,46,48. In this context, actions encompass various climate change mitigative and adaptive actions. Eight items were included and measured a 7-point Likert scale (1 = strongly disagree, 7 = strongly agree). Internal consistency was high (Cronbach’s α = 0.89).

  1. 1.

    I am willing to use public transport.

  2. 2.

    I am willing to walk or bike to destinations that are close by.

  3. 3.

    I am willing to save energy such water and electricity.

  4. 4.

    I am willing to consume less (e.g. not buying unnecessary items).

  5. 5.

    I am willing to reduce waste.

  6. 6.

    I am willing to get an education about climate change.

  7. 7.

    I am willing to join climate action campaigns and events.

  8. 8.

    I am willing to push the government to take climate action.

  • Self-efficacy refers to an individual’s belief in their capacity to take meaningful action and achieve specific outcomes in mitigating or adapting to climate change49. It is a critical psychological construct that positively influences pro-environmental behaviours50. Four items were included and measured a 7-point Likert scale (1 = strongly disagree, 7 = strongly agree). Internal consistency was high (Cronbach’s α = 0.91).

    1. 1.

      I am confident in my ability to take climate action

    2. 2.

      I am confident in my ability to change my behaviour to reduce the impact of climate change.

    3. 3.

      I am confident in my ability to engage others in climate action.

    4. 4.

      I am confident in my ability to stay informed about climate change and solutions.

Background variables

Background variables included in the study were age (in years), grade (10, 11, and 12), gender (male, female, other), religion (Islam, Christianity, Catholicism, Hinduism, Buddhism, and Confucianism), ethnicity (open text), location (based on the school municipality), past experience (yes, no), and personal values. For the past experience, I asked “Have you felt or experienced the impacts of climate change?” If yes, respondents had the chance to explain their experience in detail.

Personal values, as guiding principles in people’s lives, play a crucial role in shaping thoughts and actions, often unconsciously34. According to Steg et al.35, personal values can be categorised into four types: biospheric, altruistic, egoistic, and hedonic. Biospheric values involve a concern for the environment, while altruistic values focus on the welfare and wellbeing of others. Egoistic values pertain to personal resources and self-interest, and hedonic values emphasise pleasure and comfort. Adapted to their work, personal values included four items on biospheric values, four items on altruistic values, five items on egoistic values, and three items on hedonic values. They were measured using 7-point Likert scale (1 = strongly disagree, 7 = strongly agree).

Biospheric values (Cronbach’s α = 0.91)

  1. 1.

    It is important to me to prevent environmental pollution.

  2. 2.

    It is important to me to protect the environment.

  3. 3.

    It is important to me to respect nature.

  4. 4.

    It is important to me to be in unity with nature.

Altruistic values (Cronbach’s α = 0.75)

  1. 1.

    It is important to me that every person has equal opportunities.

  2. 2.

    It is important to me that every person is treated justly.

  3. 3.

    It is important to me that there is no war or conflict.

  4. 4.

    It is important to me to be helpful to others.

Egoistic values (Cronbach’s α = 0.71)

  1. 1.

    It is important to me to have control over others’ actions.

  2. 2.

    It is important to me to have authority over others.

  3. 3.

    It is important to me to be influential.

  4. 4.

    It is important to me to have money and possessions.

  5. 5.

    It is important to me to work hard and be ambitious.

Hedonic values (Cronbach’s α = 0.84)

  1. 1.

    It is important to me to have fun.

  2. 2.

    It is important to me to enjoy life’s pleasures.

  3. 3.

    It is important to me to do things I enjoy.

Data analyses

This study employs a multi-method analytical approach to investigate climate change conceptions among Indonesian adolescents. The analysis integrates descriptive statistics, reliability assessment, multiple linear regression, and cluster analysis, enabling both statistical relationship identification and the exploration of distinct student subgroups. Descriptive statistics—including trends (positive-neutral-negative), means, standard deviations, and frequencies—were calculated using Microsoft Excel to summarise the key characteristics of the dataset. Prior to the analysis, internal consistency and reliability of the constructs were calculated by Cronbach’s α with SPSS Statistics (version 30.0).

To examine the relationships between background variables and students’ conceptions of climate change, multiple linear regression analysis was conducted in SPSS Statistics (version 30.0) following a stepwise approach. Climate change conceptions were operationalised as a composite score, derived by combining the points of all variables. Dummy variables were created for categorical independent variables with more than two levels. Background variables included age, gender, religion, location, personal values, and past experiences, while climate change conceptions were operationalised as a composite score, combining responses across cognitive, affective, and behavioural dimensions. Dummy variables were created for categorical independent variables with more than two levels to ensure proper inclusion in the regression model.

To further explore patterns within the data, cluster analysis was conducted to categorise students based on their conceptions of climate change. This classification facilitates a learner-centred approach, allowing for targeted educational interventions based on students’ conceptual similarities and differences4. A two-step approach was implemented: first, hierarchical cluster analysis using Ward’s method was conducted to determine the optimal number of clusters. Subsequently, K-means cluster analysis was performed in SPSS Statistics (version 30.0) to classify students into distinct groups. Finally, JMP (version 16.0) was used for advanced data visualisation, enabling clear presentation of the clusters.