Climate change induced by human activities has reached unprecedented levels, with its associated adverse ecological and societal effects intensifying annually. This trend renders the global task of addressing climate change increasingly urgent (WMO, 2025; IPCC, 2023). The latest decision of the United Nations Framework Convention on Climate Change (UNFCCC, 2023) in 2023 has reinforced the necessity for global coordinated responses and calls for systemic actions to mitigate climate risks across nations. Climate literacy serves as a cornerstone for mitigating the impacts of climate change (Kumar et al., 2023). Specifically, climate literacy refers to an understanding of how the climate system works, how human actions influence the climate, and how climate change affects humans and other components of the Earth system (USGCRP, 2024). Individuals who are climate-literate understand the fundamental principles of the Earth’s climate system and the available options for responding to anthropogenic climate change; they are able to critically evaluate information related to climate change and know where to find reliable sources of such information. In addition, they can respond to climate anxiety with constructive attitudes, communicate about climate change accurately and effectively, and make informed decisions related to climate issues. Climate-literate individuals are also able to integrate cognition, emotions, attitudes, values, and actions to address climate change (USGCRP, 2024; UNESCO, 2024). Given the urgency of climate change and the individual, societal, and ecological significance of climate literacy, the United Nations Educational, Scientific and Cultural Organization (UNESCO) advocates for the integration of climate change issues into school curricula to foster students’ climate literacy (UNESCO, 2023). Among various school subjects, science education has unique value in shaping students’ climate literacy, as climate literacy must be rooted in a solid scientific foundation (Choi et al., 2021). Climate change is a complex socio-scientific issue (Choi et al., 2021); mitigating and adapting to it requires not only scientific understanding and serious attitudes towards the climate problem but also the coordination of various stakeholders based on this foresight to foster collective action. However, due to a lack of scientific literacy and the inappropriate spread of pseudoscience, climate change skepticism has become rampant, hindering societal progress towards climate action (Sharma, 2012; Dunlap, 2013). This implies that while advocating for widespread and informed climate action, we must also publicly challenge climate change skepticism. Winning this debate relies on rigorous reasoning based on scientific evidence regarding climate change. This is precisely the strength of science education. It can ground the cultivation of students’ climate literacy in the development of their socio-scientific reasoning skills, allowing the two to mutually reinforce each other. Specifically, climate change provides a vivid topic and context for students’ socio-scientific reasoning. The complex thinking, comprehensive perspectives, continuous inquiry, and critical spirit emphasized by socio-scientific reasoning (Romine et al., 2020), when combined with climate change instruction, will deepen students’ understanding of the complexity of climate issues and enable them to make responsible decisions and informed actions based on synthesizing diverse existing perspectives and attitudes towards climate problems. Therefore, the relationship between science education and students’ climate literacy is receiving increasing research attention.

Research demonstrates that science education exerts multifaceted positive effects on students’ climate literacy. First, multiple studies utilizing pre- and post-standardized tests have confirmed a significant enhancement in students’ understanding of core climate change concepts following a period of science education (McNeill and Vaughn, 2012; Bodzin and Fu, 2014). For instance, the study by Bodzin and Fu (2014) revealed that after a 20-day geospatial curriculum, middle school students’ scores on assessments of atmosphere, greenhouse effect, and climate systems significantly increased from a pre-test average of 6.23 to a post-test average of 9.47, indicating that the science education intervention substantially strengthened students’ conceptual understanding of climate change. Second, several studies have found that systematically integrating climate change content into science curricula can effectively foster students’ positive attitudes and sense of responsibility towards climate issues (Kumar et al., 2023; Clark, 2024; Karpudewan et al., 2015). For example, Kumar et al. (2023), within the “Heat-Cool Initiative,” observed that secondary students participating in a thermal imaging curriculum not only showed significantly improved test scores but also underwent an attitudinal shift. They progressed from perceiving climate change as a distant and abstract threat to believing it is personally relevant and that they can effect positive change through concrete actions, such as opting for green infrastructure. Finally, science education also contributes to the transformation of students’ climate actions (Deisenrieder et al., 2020; Choi et al., 2021). Illustratively, Choi et al. (2021), through the implementation of an SSI-STEAM project, documented improvements in both the level of student engagement in climate actions and the breadth of action domains covered.

Although the aforementioned studies affirm the positive contribution of science education to cultivating students’ climate literacy, multiple empirical studies reveal significant imbalances in its effectiveness across cognitive, affective, and action dimensions. First, science education’s promotion of climate cognition has clear limitations. On one hand, the effectiveness of science curricula in promoting climate cognition shows strong topic dependence, often failing to reach significance levels in specific areas such as wetland conservation, solar activity, or adaptation strategies (Lombardi et al., 2013, 2018; Clark, 2024). On the other hand, even well-designed science instruction often struggles to systematically reconstruct students’ deeply ingrained pre-scientific conceptions, leading to persistent confusion and misconceptions about core mechanisms like the greenhouse effect (Porter et al., 2012; Bush et al., 2019). This unstable cognitive foundation weakens the translation of knowledge into affective identification. Research shows that science curricula, even when they lead to gains in climate knowledge, may not necessarily improve students’ attitudes towards the environment (Karpudewan et al., 2015) or personal concern about climate issues (Bush et al., 2019); even courses designed to strengthen beliefs and resilience have limited effects on alleviating psychological anxiety (Gupta et al., 2025). This barrier in translating cognition to attitude ultimately makes it difficult for science education to effectively promote behavioral change in students. Extensive evidence shows that mere mastery of scientific content is insufficient to support students’ willingness to act on complex climate challenges (Igboanugo and Naiho, 2024), nor does it drive students to take practical environmental actions (McNeill and Vaughn, 2012; Clark, 2024). Furthermore, the realization of individual agency remains constrained by social norms and infrastructure beyond the curriculum system, often causing the development of action to lag significantly behind the acquisition of knowledge (Gupta et al., 2025).

Given the divergent findings among existing individual studies, it is necessary to employ meta-analysis to synthesize existing research conclusions and holistically explore the impact of science education on students’ climate literacy. Current relevant meta-analyses have primarily investigated general factors influencing climate literacy (Bergquist et al., 2023) or assessed the overall impact of environmental education on environmental literacy (Van De Wetering et al., 2022). Although Aeschbach et al. (2025) recently published a meta-analysis on the effectiveness of climate change education (CCE), their aim was to evaluate the effects of CCE in a broad sense. The literature they analyzed thematically encompassed environmental education and education for sustainable development extensively, without a specific focus on cultivating students’ climate literacy. In terms of implementation, it included both thematic education and multidisciplinary integrated instruction, thereby obscuring the distinct contribution of science curricula to climate change education. Consequently, there is currently a lack of meta-analytic research specifically targeting the effectiveness of implementing climate change education within school science education—that is, on the efficacy of cultivating students’ climate literacy.

The present study aims to employ a meta-analysis to examine whether school science education serves as an effective vehicle for developing students’ climate literacy, thereby clarifying the unique contribution of science education to climate change education. Furthermore, it seeks to investigate the factors that influence the effectiveness of school science education in fostering students’ climate literacy, to identify the specific conditions under which science education can effectively fulfill this role. This original research is expected to provide an evidence-based foundation and practical insights for enhancing school science curricula integrated with climate change education and for improving civic climate literacy. To achieve these research objectives, the following four research questions are proposed:

This study strictly adheres to the PRISMA guidelines (Page et al., 2021). The research process primarily comprises the following six steps: (1) systematic literature search and selection, (2) data extraction and coding, (3) effect size calculation, (4) effect size pooling based on a three-level model, (5) moderator analysis, and (6) discussion of findings.

In reviewing previous literature, although the topic of school science education for cultivating students’ climate literacy has received extensive attention from numerous scholars, there is a lack of a comprehensive analysis of the effects of school science education in developing students’ climate literacy. To gain a more comprehensive understanding of the effects of school science education on students’ climate literacy, this study conducted a meta-analysis, which not only verified the positive effects of school science education on students’ climate literacy but also verified which factors affected this nurturing effect under different conditions.

The 60 effect sizes from the 33 included articles produced a significantly positive impact, confirming the effectiveness of school science education in enhancing students’ climate literacy. This result resonates with findings from multiple climate education meta-analyses (Aeschbach et al., 2025; Bottin et al., 2023; van de Wetering et al., 2022). Although previous studies either focused on the role of environmental education on environmental literacy, directly focused on climate change education, or discussed climate literacy in generalized educational contexts without specifically targeting the impact of school science education on students’ climate literacy, they all reached a consistent conclusion: providing students with relevant environmental or climate change education can have a significantly positive impact on their climate literacy. Simultaneously, this study found significant effect sizes for the cognition and attitude dimensions, with school science education being more effective in promoting students’ climate change cognition. The effect size for the action dimension did not reach significance. This further validates the common pattern in climate literacy cultivation of “cognition easy to improve, action difficult to translate” (Larrain et al., 2024; Aeschbach et al., 2025; Van de Wetering et al., 2022). It is important to note that since the action dimension only included 2 studies and 3 effect sizes, there may be issues of insufficient statistical power. This suggests that conclusions regarding the action dimension are highly exploratory and should be interpreted with caution.

This study found high heterogeneity among the included effect sizes. Through three-level variance decomposition, it was found that within-study variation (Level 2) accounted for 53.3% of the total, indicating that differences in literacy dimensions may be a core source of heterogeneity. In this study, the high effect size for school science education promoting climate literacy was primarily contributed by the cognition dimension. This aligns with the phenomenon noted by UNESCO that climate cognition learning receives widespread emphasis in teaching practice (UNESCO, 2019). Furthermore, SDG 4.7 lists “ensuring that all learners acquire the knowledge and skills needed to promote sustainable development” as a core global education task. Formal science education, with its structured, systematic curriculum framework, becomes a key vehicle for achieving this goal, capable of producing significant knowledge gains in a short time (Sánchez-Almodóvar et al., 2022). From a measurement perspective, climate change cognition (e.g., climate knowledge) has clear standardized assessment tools, and its improvement is easily quantifiable (Aeschbach et al., 2025; Bottin et al., 2023), which naturally manifests as significant effect sizes in meta-analysis. Moreover, current educational practice itself centers on knowledge transmission, with curriculum design, teaching implementation, and resource allocation tilting towards cognitive cultivation, further strengthening intervention effectiveness at this level (Aeschbach et al., 2025; Bottin et al., 2023). More importantly, the rigorous mechanistic deconstruction of climate issues by science education endows cognitive learning with irreplaceable depth. For example, chemistry disciplines analyze carbon cycle mechanisms from a molecular perspective, and biology disciplines interpret ecosystem feedback mechanisms to climate change (Mahaffy et al., 2024). This in-depth analysis of scientific principles allows students to establish rigorous causal logic, making improvements at the cognitive level appear particularly solid.

Through the three-level meta-analysis model, we found that 45% of the total heterogeneity also stemmed from between-group differences. Moderator effect analysis showed that discipline was a source of this heterogeneity, while teaching strategy, educational level, and intervention duration did not significantly influence the research results.

Our analysis confirmed that discipline significantly moderates the impact of school science education on students’ climate literacy, which aligns with Onuoha et al. (2021). However, a nuanced discrepancy exists in the ranking of disciplinary effects. Onuoha et al. (2021) identified geography as the most effective discipline for enhancing students’ climate awareness among the four disciplines examined, followed by chemistry, biology, and science. In contrast, our meta-analysis revealed that chemistry yielded the largest effect size, followed by biology, geography, Earth science, and science. This discrepancy may stem from methodological differences. Onuoha et al. (2021)’ study was a single empirical research in which the authors used correlation analysis to reveal associations between disciplines and climate awareness. Our study employed meta-analysis, drawing conclusions based on the synthesis of multiple empirical studies. The reason chemistry produced the most significant moderating effect may lie in its ability to explain the causes of climate change at the micro-mechanistic level (e.g., molecular structures of greenhouse gases), providing students with a solid conceptual knowledge base (Versprille and Towns, 2015). Simultaneously, its systematic exposition of material cycles (e.g., the carbon cycle) can link individual actions to global impacts, thereby effectively cultivating students’ systems thinking skills (Mahaffy et al., 2024). Furthermore, chemistry education can make abstract solutions concrete and credible by explaining the chemical principles behind teaching content such as carbon capture and new energy conversion, making students aware of the realistic possibility of human actions affecting the climate system, thus effectively promoting the transformation of cognition into action intention (Mahaffy et al., 2024; Clark, 2024). In contrast, the moderating effect of general science was weakest, possibly because science courses have comprehensive characteristics, and when learning interdisciplinary topics like climate systems, which involve abstract models, students are prone to feelings of boredom, abstraction, and difficulty. This can diminish the effectiveness of their climate learning (Winarno et al., 2020). It should be noted that because the number of studies and effect sizes for chemistry, biology, and geography are far lower than those for earth science and general science, we need to treat the discipline ranking cautiously, viewing it as an exploratory finding. Nevertheless, despite differences in effect sizes, this meta-analysis shows that all science disciplines can significantly enhance students’ climate literacy. Therefore, continuing discipline-based science teaching related to climate change issues is worthwhile. Furthermore, some studies point out that relying solely on single-discipline approaches may not yield the best results (Salinas et al., 2022), so it is also necessary to explore multidisciplinary and interdisciplinary projects alongside integration within individual subjects to optimize climate change education.

Instructional strategy was not a significant moderator variable, indicating that the effectiveness of school science education in fostering students’ climate literacy is not contingent on whether digital technologies are used. On the one hand, this may be because climate change is a multi-layered issue that spans past, present, and future temporally and encompasses local, regional, and global scales spatially (USGCRP, 2024). Whether science classrooms connect students’ everyday experiences through traditional instructional approaches or simulate temperature trends using digital tools, students can experience strong contextual engagement (Igboanugo and Naiho, 2024; Mostacedo-Marasovic et al., 2023). On the other hand, this may reflect the fact that the effectiveness of technology depends heavily on how it is pedagogically integrated rather than on its mere presence (Wu, 2024). As long as the goals, content, and methods of climate literacy instruction are coherently aligned to actively engage learners, effective outcomes can be achieved (UNESCO, 2024). Existing research also indicates that both digital and traditional instructional approaches have their respective advantages in fostering students’ environmental-related literacies. Mulders and Träg (2023) found that extended reality (XR) could impart knowledge about biodiversity and influence students’ environmental attitudes. Labianca (2021) argued that integrating Geographic Information Systems (GIS) into geography teaching helps enhance local resource awareness. Boaventura et al. (2020) revealed that science teaching centered on inquiry-based experimental activities significantly enhanced students’ cognition and explanatory ability regarding the impacts of marine climate change. The study by Restović and Bulic (2024) showed that students’ stays and participation in environmental activities in national parks could significantly influence their positive attitudes towards nature conservation and environmental knowledge. These findings offer valuable insights into how we should navigate the relationship between digital empowerment and traditional teaching in the era of educational digitization. It should be noted that while the current dichotomous approach maximizes the robustness of meta-analysis results, it may obscure the heterogeneity of effects within similar strategies (e.g., among different digital tools). Therefore, this conclusion should be regarded as exploratory, and further research is needed to deepen our understanding of the roles played by different teaching strategies.

Educational level did not significantly moderate the impact of school science education on student climate literacy. This finding is consistent with Aeschbach et al. (2025), who found that educational level did not significantly moderate the effectiveness of climate change education across multiple dimensions including knowledge, attitudes, and behavior. We believe the lack of a significant moderating effect from educational level may stem from two aspects. On one hand, existing climate competency assessment scales may not have differentiated indicators designed for different educational levels, preventing the assessment results from reflecting true differences between educational levels (Horry et al., 2023). On the other hand, students across grade levels may generally lack the core cognitive abilities required to understand climate science, as revealed by Jin et al. (2013), who found that students from fourth to twelfth grade commonly held misconceptions about carbon emission mechanisms. For instance, they generally believed that gasoline “disappears” or “turns into energy” during combustion, overlooking the reaction with oxygen that produces carbon dioxide. It should be noted that due to the severe imbalance in sample sizes across educational attainment subgroups, the current findings should be regarded as exploratory and interpreted with caution.

Intervention duration was not a significant moderator variable, suggesting that the effectiveness of school science education in fostering students’ climate literacy is not proportional to the length of time. To date, research has not reached a consensus on this issue. A meta-analysis by Aeschbach et al.’s (2025) identified duration as a significant moderator of climate change education effects, and Sabarwal et al. (2024), based on an analysis of 96 countries worldwide, found that each additional year of education increases climate awareness by 8.6%. Çalik and Wiyarsi’s (2025) meta-analysis similarly showed that intervention duration significantly moderated the effectiveness of instruction on socioscientific issues. In contrast, Van De Wetering et al. (2022) reported that duration did not significantly moderate environmental education outcomes, and Doğan et al. (2023) found that the effectiveness of the flipped classroom approach in science education was not significantly moderated by duration. The present study corroborates the latter, non-significant findings. In this study, the lack of a significant moderating effect of intervention duration may be attributed to the fact that climate literacy is not a linearly cumulative learning outcome (Aeschbach et al., 2025), and its constituent sub-dimensions do not necessarily develop synchronously (Klement et al., 2025). Knowledge- and attitude-related outcomes concerning climate change may emerge within relatively short periods, while extending the duration of instruction may primarily serve to consolidate and deepen learning rather than proportionally amplify effect sizes. Thus, interventions lasting <1 month may already be sufficient to “initiate” climate literacy development (Pacini et al., 2025; Lombardi et al., 2013), and longer durations do not necessarily yield substantially higher average effects. For example, Pacini et al. (2025) employed only a four-hour instructional intervention and found that students’ environmental knowledge remained significantly higher than pre-intervention levels 4 weeks later. Lombardi et al. (2013) used a 90-min intervention and observed that students’ understanding of anthropogenic climate change remained stable 6 months after the intervention. Another possible explanation is that intervention quality matters more than duration. Short-term interventions (e.g., < 1 month) may benefit from focused and intensive climate change themes due to time constraints, while long-term interventions (e.g., more than 6 months) may suffer from unfocused themes and diluted content. It should also be noted that the number of studies and sample sizes varied substantially across duration subgroups (e.g., 16 studies in the “< 1 month” group versus only 7 studies in the “1–6 months” group), which may have limited the statistical power to detect true differences. Therefore, these findings should be interpreted with caution. Considering the potential limitations of the present study and the inconsistencies in existing research, the moderating role of intervention duration remains an important topic for further investigation.

As a novel attempt to comprehensively analyze the impact of school science education on student climate literacy, this meta-analysis found that school science education exerts a significantly large and positive overall effect on student climate literacy. Sub-dimension analysis showed that school science education is most prominent in enhancing students’ climate change cognition, followed by attitude, while action did not reach significance. This study also found that discipline effectively moderates the impact of school science education on students’ climate literacy, while teaching strategy, educational level, and intervention duration are not key moderating factors.

Based on these findings, this study proposes four practical recommendations. First, systems thinking should be employed in top-level design to deeply integrate climate change topics into school science education. On one hand, integration should be organic and aligned with disciplinary characteristics. This meta-analysis confirms that all science disciplines contribute to developing student climate literacy, albeit with varying effectiveness. This suggests a need to both explore effective educational strategies grounded in specific disciplinary features. On the other hand, interdisciplinary collaborative projects should be developed. These should leverage the core strength of disciplines like chemistry in explaining climate change mechanisms, while synergistically utilizing the complementary advantages of geography, biology, and earth science in areas such as spatial distribution, ecosystems, and scientific observation. This approach transcends the limitations of single-discipline instruction, holistically enhancing students’ systemic understanding of and capacity to respond to climate issues. Second, a balanced and comprehensive cultivation of students’ climate change cognition, attitudes, and action should be emphasized. This meta-analysis reveals that cognitive outcomes yield the strongest effects, followed by attitudes, while actions remain insignificant. Given the volume of studies we included, the number of cognitive-related findings also far exceeds those concerning attitudes and actions. This suggests that current science education may exhibit an overemphasis on cognitive development when cultivating climate literacy. Therefore, while maintaining the cognitive strengths of science education, greater attention must be directed toward nurturing students’ attitudes, emotions, values, and actions regarding climate issues. Third, teaching strategies should be selected judiciously based on instructional objectives. Although teaching strategy was not a significant moderator variable, this meta-analysis found that both traditional teaching and science teaching using digital tools can effectively enhance students’ climate literacy. This insight suggests that when designing instruction integrating climate change topics, science teachers should not hold preconceived notions about any particular teaching strategy. Rather, they should recognize that any strategy serves as a means to achieve instructional goals and content. Whether to adopt traditional methods or integrate digital technology should be primarily determined by the goal of developing climate literacy. For example, cultivating ecological care is ideally achieved through nature experiences and field inquiry. Cultivating understanding of phenomena like surface temperature rise can be effectively addressed through computer climate modeling. Fourth, when designing and implementing climate change education programs, science teachers should shift their focus from increasing intervention duration to optimizing intervention design. This meta-analysis found that the effectiveness of science education in cultivating students’ climate literacy is not necessarily linked to intervention duration. This suggests that on one hand, the effects of short-term interventions should be valued; on the other hand, more effective long-term intervention strategies should be explored to avoid merely increasing time without improving effectiveness.

This study has several limitations. First, this meta-analysis only included studies published in English. However, given the global nature of climate change education, this may lead to an overestimation of the effect estimate. Furthermore, this study only included experimental and quasi-experimental research, potentially systematically excluding high-quality longitudinal or mixed-methods studies that more authentically capture behavioral changes. This design limitation may bias results towards cognitive gains. Future research should include studies in multiple languages and various research types to understand this phenomenon more comprehensively. Second, although this study’s three-dimensional division of climate literacy has a basis in the literature, it must be acknowledged that this classification has inherent limitations. For example, “cognition” may blur the distinction between knowledge and perception, and “attitude” may blur the distinction between values and emotions. As research on each sub-dimension increases, future meta-analyses need to classify more meticulously and cautiously. In this study, the action dimension had only three effect sizes, which not only made comparisons across the three dimensions difficult but also limited further moderator analysis by sub-dimension. Therefore, we recommend that future research increase attention to climate action. Third, due to the limitations of the included literature, this study only examined discipline, educational level, intervention duration, and teaching strategy as moderator variables, unable to test the potential influence of other relevant factors like teachers on students. Moreover, the moderator effect analysis for teaching strategy lacked more detailed analysis. This shortcoming stems from an insufficient number of studies on specific strategies, preventing us from establishing independent categories for effective analysis. Future research should explore the potential impact of other moderator variables on the effectiveness of school science education in cultivating students’ climate literacy more broadly and strive to focus more on teaching strategies for targeted exploration to better elucidate how various strategies differentially promote students’ climate literacy. Finally, given the significant heterogeneity among studies, the conclusions of this meta-analysis should be interpreted cautiously. It is recommended that future research adopt more unified methodological standards and conduct more experiments with high homogeneity to support more robust meta-analytic conclusions.