Content area
Scientific creativity is a critical component of science education, promoting innovative thinking and the ability to address complex, real-world challenges. Despite its importance, quantitative syntheses specifically examining the effects of scientific problem-solving approaches on fostering scientific creativity are scarce. This meta-analysis addresses this gap by evaluating the impact of inquiry-based learning, problem-based learning, project-based learning, and STEM contexts on students’ scientific creativity. Using the PRISMA protocol, 19 studies were analyzed, revealing that PjBL and PBL produced the most substantial effects, with large effect sizes (g = 2.10 and g = 1.49, respectively), while STEM contexts and IBL demonstrated moderate positive impacts. These findings are interpreted within the scientific method, emphasizing critical stages such as problem definition, hypothesis formulation, experimentation, data analysis, and evaluation. The results suggest that these approaches effectively nurture scientific creativity by engaging students in interdisciplinary, real-world problem contexts that demand innovative solutions. However, the limited number of studies and potential publication bias present challenges to broader generalization. Future research should expand the evidence base and explore the integration of these approaches across diverse educational settings. This study contributes to the growing body of evidence supporting the role of scientific problem-solving approaches in cultivating scientific creativity, offering valuable insights for educators and curriculum developers.
Introduction
Scientific creativity has emerged as a critical competency in the 21st century, particularly within science education, where it drives innovation, discovery, and the ability to address complex, real-world challenges (Kim et al., 2019; Fernandez et al., 2024). It goes beyond routine problem-solving by encouraging students to think divergently, generate original ideas, and apply scientific knowledge in novel ways (Stehle & Peters-Burton, 2019). As modern societies become increasingly dependent on scientific and technological advancements, equipping students with creative scientific thinking skills has become essential for addressing global challenges and advancing knowledge (Acut, 2022; Abulibdeh et al., 2024). However, fostering scientific creativity remains a significant pedagogical challenge, requiring approaches that transcend traditional, lecture-based teaching methods (Ammar et al., 2024).
Among the various strategies for enhancing scientific creativity, problem-solving approaches like inquiry-based learning (IBL), problem-based learning (PBL), project-based learning (PjBL), and STEM/STEAM contexts have gained prominence (Demirhan & Sahin, 2021; Gunawan et al., 2018b; Siew & Ambo, 2020). These approaches are rooted in constructivist theories that emphasize active learning, collaboration, and real-world application, aligning closely with the scientific method’s core stages, to wit: problem definition, hypothesis formulation, experimentation, data analysis, and evaluation (Yildiz & Yildiz, 2021). For instance, IBL encourages students to pose questions and design investigations, PBL focuses on solving authentic, open-ended problems, and PjBL emphasizes creating tangible products as a means of understanding scientific concepts. STEM contexts further extend these approaches by integrating interdisciplinary content, fostering a holistic understanding of scientific phenomena (Çalışıcı & Benzer, 2021).
Despite the potential of these approaches to cultivate scientific creativity, their effectiveness remains underexplored in the literature. While previous meta-analyses, such as the one by Bi et al. (2020), have examined the impact of various teaching strategies on scientific creativity, these reviews often include a broad range of interventions, such as general problem-solving, collaborative learning, and scientific reasoning, without specifically focusing on the distinct contributions of IBL, PBL, PjBL, and STEM contexts. Moreover, many of these studies are outdated, with the most recent analyses incorporating data only up to 2019, underscoring the need for a more current and focused synthesis of evidence.
Purpose of the study
This meta-analysis aims to fill this gap by providing a comprehensive, up-to-date synthesis of the effectiveness of scientific problem-solving approaches in fostering scientific creativity. Specifically, this study seeks to:
Determine the overall effectiveness of IBL, PBL, PjBL, and STEM contexts in enhancing students’ scientific creativity;
Conduct subgroup analyses to identify potential moderators influencing these effects;
Assess potential publication bias in the included studies; and.
Synthesize findings in relation to the scientific method, highlighting critical stages such as problem definition, experimentation, and evaluation.
These objectives aim to provide educators and curriculum developers with evidence-based insights into the most effective strategies for cultivating scientific creativity, ultimately contributing to a deeper understanding of innovative science education practices.
Literature review
Role of creativity and scientific creativity in science education
Creativity has emerged as a critical educational outcome in the 21st century, driven by the need for innovative thinking in an increasingly complex, digitized, and interconnected world (Long et al., 2022). It is recognized as one of the core competencies within the “learning and innovation skills” framework, distinguishing students prepared for a rapidly changing, globally connected, and technologically advanced society (P21’s Framework for 21st Century Learning, 2019; Hernández-Torrano & Ibrayeva, 2020). According to the Revised Bloom’s Taxonomy, creativity represents the highest level of cognitive complexity, requiring students to synthesize knowledge in novel and meaningful ways (Armstrong, 2010). This capacity for original thought is essential for addressing real-world problems, fostering innovation, and driving scientific and technological progress (Kaufman & Sternberg, 2019; Panergayo & Prudente, 2024).
Scientific creativity, a specialized subset of creativity, is defined as the ability to generate novel, valuable, and scientifically relevant ideas (Simonton, 2012). It is crucial in developing critical scientific skills, including hypothesis generation, experimental design, data analysis, and interpretation (Xu et al., 2024). Scientific creativity not only involves the creation of new scientific knowledge but also the ability to approach problems from multiple perspectives, adapt to new information, and make connections across disciplines (Ward, 2007). These skills are increasingly important as scientific challenges become more complex and interdisciplinary, requiring innovative solutions and flexible thinking (Sandri, 2013).
Assessing scientific creativity
Several models have been developed to assess and understand scientific creativity. One prominent framework is the Scientific Structure Creativity Model (SSCM) developed by Hu and Adey (2002), which conceptualizes scientific creativity across three dimensions:
Product Dimension – Includes technical products, scientific knowledge, scientific phenomena, and problems, emphasizing tangible outputs of scientific thought.
Process Dimension – Focuses on the cognitive processes involved in creativity, such as imagination, hypothesis generation, and innovative thinking.
Trait Dimension – Draws on Torrance’s (1990) primary aspects of creativity, including fluency (the number of relevant ideas), flexibility (the ability to adapt and shift perspectives), and originality (the uniqueness of responses).
This model provides a comprehensive framework for evaluating and nurturing scientific creativity in educational settings, capturing both the creative process and the resulting outcomes. It highlights the importance of fostering not just factual knowledge but also the innovative thinking required for scientific discovery (Hu & Adey, 2002).
Scientific method as a foundation for scientific creativity
The scientific method provides the foundational framework for cultivating scientific creativity. It emphasizes systematic inquiry, empirical evidence, and critical reasoning, aligning closely with the cognitive processes involved in creative scientific work (Popper, 1959; Vo & Simmie, 2024). While often perceived as a rigid, linear process, the scientific method inherently involves creative thinking at each stage, including:
Problem identification and question formulation– Identifying meaningful scientific problems requires creative insight, as scientists must often frame questions in novel ways to explore uncharted areas of knowledge (Runco, 2007). For instance, Charles Darwin’s innovative interpretation of natural observations led to the groundbreaking theory of evolution.
Hypothesis development – Crafting testable hypotheses demands both scientific reasoning and imaginative thinking, aligning closely with the fluency and originality components of the SSCM (Hu & Adey, 2002).
Experimental design – Designing effective experiments involves innovative problem-solving, anticipating challenges, and creatively controlling variables to generate reliable data (Gopnik, 2012).
Data analysis and interpretation – Analyzing complex data requires flexible thinking to identify patterns and connections that are not immediately apparent, a critical aspect of scientific creativity (Ward, 2007).
Evaluation and communication – Effectively presenting scientific findings involves not only technical precision but also creative storytelling to make complex ideas accessible to diverse audiences (Simonton, 2012).
Pedagogical approaches for fostering scientific creativity
The scientific method provides a foundational structure for various pedagogical approaches that aim to foster scientific creativity. These approaches emphasize inquiry, evidence-based reasoning, and iterative experimentation, aligning closely with the cognitive processes required for scientific innovation:
PBL – PBL immerses students in real-world scientific problems, guiding them through the stages of the scientific method. This includes defining the problem, generating hypotheses, designing and conducting investigations, and drawing evidence-based conclusions (Hmelo-Silver, 2004). This approach mirrors the iterative and exploratory nature of scientific inquiry, fostering creativity by encouraging students to think critically and flexibly in response to complex challenges.
PjBL – PjBL emphasizes long-term, interdisciplinary projects that require sustained inquiry, creative problem-solving, and iterative refinement. Students engage in processes similar to the scientific method, including problem identification, data collection, analysis, and reflection, promoting both scientific creativity and deeper conceptual understanding (Larmer & Mergendoller, 2010).
IBL – IBL encourages open-ended questioning, systematic experimentation, and evidence-based reasoning, fostering scientific creativity by promoting curiosity, critical thinking, and a willingness to explore unknowns (Panjaitan & Siagian, 2020). This approach closely mirrors the stages of the scientific method, supporting students in developing the cognitive skills needed for scientific innovation.
STEM/STEAM contexts – STEM (Science, Technology, Engineering, and Mathematics) and STEAM (adding Arts) education integrate hands-on, investigative learning that requires students to apply scientific reasoning and creative problem-solving to interdisciplinary challenges (Tran et al., 2021). While STEM is often characterized as an interdisciplinary framework, it fundamentally relies on the scientific method as a structure for inquiry, experimentation, and innovation (Zhong et al., 2022).
These approaches collectively support the development of students’ creative skills, preparing them to address scientific challenges with fresh perspectives. They not only engage students in rigorous scientific inquiry but also foster the innovative thinking necessary for scientific discovery and technological advancement (Ward, 2007). Grounded in the principles of the scientific method, these pedagogical strategies provide structured yet flexible frameworks that enhance scientific creativity, critical thinking, and problem-solving abilities.
Creating a supportive environment for scientific creativity
Fostering scientific creativity requires a supportive classroom environment where exploration, collaboration, and innovative thinking are actively encouraged (Stolaki et al., 2023). Educators play a crucial role by promoting a growth mindset that values the creative process over rote memorization and standardized testing (Ward, 2007). This involves creating learning spaces where students feel safe to take intellectual risks, experiment with new ideas, and learn from failure, recognizing that creativity often emerges through trial and error.
Collaboration and open communication further enhance creativity by allowing students to share diverse perspectives and build on each other’s ideas (Sandri, 2013). Group work and peer discussions foster a culture where students can critique constructively and refine their thinking, leading to unique solutions to complex scientific challenges. Such interactions help students develop not only creative skills but also important social and cognitive competencies.
Hence, fostering scientific creativity prepares students for lifelong learning and innovation in a rapidly evolving world. Encouraging iterative experimentation and real-world problem-solving nurtures flexible thinking and resilience, equipping students to navigate scientific challenges with confidence and originality. This supportive environment is essential for developing the creative capacities that modern science education aims to cultivate.
Methodology
Research design
This study adopts a meta-analysis research model, a quantitative method that integrates and synthesizes findings from multiple independent studies to draw broader, evidence-based conclusions (Borenstein et al., 2009; Schmidt & Hunter, 2015). Meta-analysis is particularly effective for resolving conflicting results across studies by accounting for variations in sample sizes, measurement tools, and study designs, thereby enhancing the reliability and generalizability of the findings.
Given the study’s focus on scientific creativity, the meta-analysis specifically examines the effects of various scientific problem-solving approaches, including PjBL, PBL, IBL, and STEM contexts. These approaches were selected for their alignment with the scientific method and their potential to foster scientific creativity through authentic, inquiry-driven learning experiences.
To ensure the robustness and reliability of the findings, subgroup analyses and publication bias assessments were conducted, providing deeper insights into the conditions under which these strategies are most impactful. This approach not only strengthens the validity of the conclusions but also offers practical guidance for educators and curriculum developers seeking to promote scientific creativity.
Eligibility criteria
This meta-analysis focuses on studies that investigate the effectiveness of scientific problem-solving approaches, including PjBL, PBL, IBL, and STEM/STEAM contexts, in fostering students’ scientific creativity within science education. To ensure the rigor and comparability of the included studies, the following eligibility criteria, as outlined in Table 1, were established:
Table 1. Eligibility criteria for inclusion and exclusion of studies
Criteria Type | Details |
|---|---|
Inclusion Criteria | Study Design: Only empirical studies employing true-experimental or quasi-experimental designs with control and experimental groups were included. These designs needed to establish clear cause-and-effect relationships, using equivalent or nonequivalent pretest-posttest research designs with two or more comparable groups. |
Time Frame: Studies conducted from 2000 to 2023 were considered, reflecting recent advancements in scientific creativity research. | |
Intervention Focus: The research must introduce a specific instructional intervention aimed at fostering scientific creativity, grounded in problem-solving approaches and aligned with the scientific method. | |
Statistical Reporting: Studies were required to provide comprehensive statistical data, including mean pretest-posttest scores, standard deviations, sample sizes, t-values, f-values, or actual effect sizes, to enable accurate effect size calculation. | |
Exclusion Criteria | Studies published in languages other than English. |
Studies that primarily addressed non-creativity outcomes, such as motivation, conceptual understanding, engagement, or general cognitive skills, without a clear focus on scientific creativity. | |
Research focusing exclusively on gifted or special education populations, as these contexts may not be directly comparable to general science education. | |
Studies that did not use a science learning environment as the context for intervention. | |
Pre-experimental, correlational, or qualitative studies lacking control or comparison groups. |
These criteria were designed to ensure the robustness and reliability of the findings, allowing for a comprehensive synthesis of the impact of problem-solving approaches on scientific creativity in diverse educational settings.
Information sources
The information sources for this study include electronic databases such as Google Scholar, ERIC (Education Resources Information Center), ScienceDirect, and selected Scopus-indexed journals, focusing on science education and creativity research. Electronic databases are considered the “most fruitful” sources of information for research synthesis and meta-analysis (Cooper, 2010). Google Scholar, an open educational resource freely accessible via a web search engine, catalogues the full-text or metadata of academic literature across various publishing formats and disciplines, archiving articles, theses, books, abstracts, and court opinions from academic publishers, professional societies, online repositories, universities, and other websites.
Similarly, ERIC is a comprehensive, manageable, and accessible web-based bibliographic and full-text database of education research and information, designed for academicians, researchers, educators, policymakers, and the general public. ScienceDirect, Elsevier’s premier platform of peer-reviewed literature, provides access to a vast bibliographic database of scientific and medical publications, including research related to social science and humanities, emphasising historical context, current developments, theories, applications, and trends in education, psychology, humanities, and business. Scopus, Elsevier’s abstract and citation database, indexes publishers and peer-reviewed journals in top-level subject fields such as life sciences, social sciences, physical sciences, and health sciences.
Search and selection strategy
The search strategy for this meta-analysis involved querying multiple academic databases, including Google Scholar, ERIC, ScienceDirect, and Scopus-indexed journals, to identify studies relevant to scientific problem-solving and creativity. Keywords included “Scientific Creativity,” “Creativity in Science,” “Science Education,” and combinations like “Science AND Creativity.” Although PBL, IBL, PjBL, and STEM/STEAM were not specifically included as search terms, this approach was deliberate to ensure a broader capture of studies explicitly addressing scientific creativity as the core outcome. Using these methods as primary keywords could have inadvertently narrowed the search, potentially excluding relevant studies that integrate these pedagogical approaches without explicitly identifying them in their titles or abstracts. Instead, these approaches were included during the screening and selection stages if they aligned with the study’s focus on fostering scientific creativity through diverse, student-centered methods.
For Google Scholar, the software tool Harzing’s Publish or Perish was used to compile relevant journal articles, with publication dates restricted to 2000 to 2023. In ERIC, advanced searches included terms like “Scientific Creativity” and “Science Education,” filtered to include only full-text articles published within the last two decades. On ScienceDirect, the search strategy emphasized research articles published between 2000 and 2023 from journals such as Procedia - Social and Behavioral Sciences, Teaching and Teacher Education, and Design Studies, particularly within the social sciences and arts and humanities subject areas.
The initial selection process involved screening article titles and abstracts to identify studies potentially relevant to the development of scientific creativity within science learning contexts. Eligible articles passing this stage were recorded and organized in Microsoft Excel, applying the established inclusion and exclusion criteria. Final study inclusion was determined through consensus among multiple reviewers. Figure 1 presents the flow of the search and selection process, following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines (Moher et al., 2009).
The initial search yielded 2,852 articles related to scientific creativity and its instructional development in science education. Many studies were excluded during initial screening due to failure to meet the required experimental design criteria. From the pool, 35 articles were selected for full-text review, of which 19 articles (comprising 21 effect sizes) met all eligibility criteria and were included in the meta-analysis. The remaining 16 articles were excluded primarily due to incomplete statistical data, pre-experimental or single-group designs, or publication in languages other than English.
[See PDF for image]
Fig. 1
Flow diagram of the literature search using PRISMA protocol
Coding procedures
The data extracted from the qualified studies were systematically coded to capture essential information for the meta-analysis. Specifically, the coding framework included the following elements: study identification, which involved recording the authors’ last names and the publication year; educational level, indicating the academic level of the participants (e.g., primary, secondary, tertiary); science domain, specifying the scientific discipline addressed in the study (e.g., physics, biology, chemistry); country of study, identifying the geographical context where the research was conducted; instructional intervention, which referred to the approaches employed (e.g., PBL, PjBL, IBL, STEM/STEAM); duration of intervention, detailing the length of time over which the intervention was implemented; learning format, specifying the mode of instruction (e.g., face-to-face, blended, online); and statistical measures, including key data such as mean, standard deviation, sample size, and effect sizes for both the experimental and control groups.
The coding procedures were carried out collaboratively by the first four authors, under the direct supervision of the fifth and sixth authors. The process involved a thorough review to resolve discrepancies, including misclassified data, duplicated entries, omitted information, and typographical errors (Acut & Antonio, 2023). Misclassified entries were corrected, duplicates were removed to ensure each data point was included only once, missing information was addressed where possible, and typographical errors were rectified to maintain the dataset’s integrity.
In cases where specific data were not reported in the included articles, the missing information was coded as “Not Reported,” given the reliance on available manuscript data without direct author contact. Future researchers are encouraged to provide complete and transparent data to enhance the comprehensiveness and reliability of subsequent meta-analyses. This meticulous approach ensured the accuracy and reliability of the final dataset, providing a robust foundation for subsequent statistical analyses.
Data analysis
The data analysis for this meta-analysis followed a rigorous multi-step approach to ensure accurate and reliable findings. Initially, diagnostics for outliers and influential cases were conducted using metrics such as externally standardized residuals, as recommended by Hedges and Olkin (1985) and Viechtbauer and Cheung (2010). Outliers identified through these methods were subsequently excluded from the analysis to enhance the precision of the effect size estimates.
Effect sizes were calculated using Meta-Essentials version 1.5, a robust statistical tool designed to facilitate the integration and synthesis of effect sizes from multiple studies (Suurmond et al., 2017; Vhan Rhee et al., 2015). Hedges’ g, a widely used metric for quantifying the standardized mean difference, was chosen as the primary effect size measure. This statistic corrects for small sample biases, making it more reliable than Cohen’s d, particularly when the total sample size for the experimental and control groups is less than 20 (Cohen, 1988; Hedges & Olkin, 1985). Hedges’ g is calculated by dividing the difference between the experimental and control group means by the pooled standard deviation, adjusting for sample size to reduce the upward bias commonly associated with smaller studies (Lin & Aloe, 2020). An effect size (g) is typically interpreted as small (≤ 0.20), medium (> 0.20 but < 0.80), or large (≥ 0.80), providing a standardized measure for comparing the relative impact of different interventions.
To further explore the heterogeneity in effect sizes across various pedagogical approaches, subgroup analyses were conducted. This step is essential for understanding which instructional methods are most effective in fostering scientific creativity. In addition to subgroup analyses, the study also included a comprehensive publication bias assessment using multiple methods, including the Begg-Mazumdar rank correlation test, Egger’s regression test, and various failsafe tests such as Rosenthal’s, Orwin’s, and Fisher’s methods (Begg & Mazumdar, 1994; Egger et al., 1997; Lin & Chu, 2017; Rosenthal, 1979; Orwin, 1983; Fisher, 1932). These analyses are critical for identifying and correcting potential publication biases, ensuring the robustness and reliability of the meta-analysis findings.
Results and discussion
Characteristics of the included studies
The characteristics of the studies included in this meta-analysis are summarized in Table 2, which captures essential information such as research designs, educational levels, participant numbers, geographic distribution, and science domains.
Table 2. Characteristics of the studies included in this meta-analysis
Authors | Year of publication | Setting | Research design | Educational Level | Participants | Science domain | |
|---|---|---|---|---|---|---|---|
Experimental | Control | ||||||
Aktamis & Ergi | 2008 | Asia | QER | Primary | 20 | 20 | Physics |
Çalışıcı & Benzer | 2021 | Asia | QER | Secondary | 22 | 22 | Biology |
Chumo | 2014 | Africa | QER | Secondary | 47 | 47 | Biology |
Demirhan & Sahin | 2021 | Asia | QER | College | 21 | 21 | Biology |
Doğan & Kahraman | 2021 | Asia | QER | Secondary | 50 | 48 | STEM |
Dyson et al. | 2016 | Asia | QER | College | 19 | 20 | Physics |
Erol et al. | 2022 | Asia | QER | Primary | 33 | 33 | STEAM |
García & Fernández | 2018 | Europe | QER | Secondary | 100 | 100 | Science |
Gunawan et al. | 2018a | Asia | QER | Secondary | 98 | 97 | Physics |
Gunawan et al. | 2018b | Asia | QER | Secondary | 32 | 31 | Physics |
Hu et al. | 2013 | Asia | QER | Secondary | 54 | 53 | Science |
Kırıcı and Bakırcı | 2021 | Asia | QER | Secondary | 35 | 29 | STEM |
Ozkan & Topsakal | 2019 | Asia | QER | Secondary | 37 | 37 | STEAM |
Sari et al. | 2019 | Asia | MMR | Not reported | 40 | 40 | Biochemistry |
Siew & Ambo | 2018 | Asia | QER | Primary | 30 | 30 | STEM |
Siew & Ambo* | 2020 | Asia | QER | Primary | 120 | 120 | STEM |
Siew et al.* | 2017 | Asia | QER | Pre-school | 72 | 72 | Science |
Tran et al. | 2021 | Asia | QER | Secondary | 31 | 31 | STEAM |
Yang et al. | 2016 | Asia | QER | Primary | 20 | 24 | Science |
Note. *Studies with 2 effect sizes
The studies collectively involved 1,756 students, reflecting a substantial dataset for analyzing the effectiveness of scientific problem-solving approaches in various educational contexts. The included studies span a wide range of publication years, with a noticeable increase in recent years, particularly from 2020 onwards, indicating growing scholarly interest in fostering scientific creativity through innovative pedagogical methods.
Geographically, the majority of these studies were conducted in Asia, with a smaller number from Europe and Africa, highlighting regional differences in educational priorities and research focus. This concentration in Asia aligns with significant investments in STEM education driven by government policies aimed at equipping students with critical 21st-century skills (Cheng, 2022; Tuyet et al., 2024).
Methodologically, most studies employed quasi-experimental designs, reflecting a strong emphasis on establishing causal relationships between scientific problem-solving interventions and creativity outcomes. This preference for quasi-experimental approaches underscores the field’s commitment to methodological rigor, while the inclusion of a few true-experimental and mixed-methods studies provides additional methodological diversity, enriching the overall analysis.
In terms of educational levels, the included studies span primary to higher education, with a notable concentration at the secondary level. This range demonstrates the versatility of scientific problem-solving pedagogies, which can be effectively implemented across different age groups and educational stages. The studies also encompass a wide array of science domains, including Physics, Biology, Chemistry, STEM, and General Science, highlighting the cross-disciplinary applicability of scientific creativity as a critical educational outcome.
Participant numbers varied considerably, ranging from as few as 19 to over 100 students per study, reflecting differences in study scope, institutional resources, and experimental design. This variation indicates that scientific problem-solving approaches can be effectively scaled for both small and large learning environments, enhancing their adaptability to diverse educational contexts (García & Fernández, 2018; Hu et al., 2013).
This broad representation of educational settings, scientific disciplines, and participant group sizes provides a comprehensive foundation for understanding the factors that influence scientific creativity in education, reinforcing the importance of context-specific approaches for fostering innovative thinking and problem-solving skills.
Effect sizes of scientific problem-solving approaches
The data presented in Table 3 provides a comprehensive overview of the overall effect size and heterogeneity analysis for studies utilizing scientific problem-solving approaches. The fixed-effects model reveals an effect size of Hedge’s g = 0.97, with a 95% confidence interval ranging from 0.87 to 1.07, indicating a substantial positive impact on students’ scientific creativity. This effect size is statistically significant (Z = 20.46, p < 0.0001), suggesting that scientific problem-solving approaches have a meaningful influence on educational outcomes across the included studies. In contrast, the random-effects model shows a slightly higher effect size of Hedge’s g = 1.33, with a 95% confidence interval from 0.80 to 1.86, reflecting the broader variability among studies and confirming the overall positive impact, albeit with some heterogeneity in effect sizes.
Table 3. Overall effect size and heterogeneity analysis
Model | Effect size and 95% confidence interval | Test of null (2-Tail) | Heterogeneity statistics | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
No. of studies | Hedge’s g | Std error | Variance | Lower limit | Upper limit | Z-value | p-value | Q-value | df (Q) | p-value | I2 | |
Fixed | 21 | 0.97 | 0.05 | 0.07 | 0.87 | 1.07 | 20.46 | 0.000* | 310.62 | 20 | 0.000* | 93.56 |
Random | 21 | 1.33 | 0.26 | 0.31 | 0.80 | 1.86 | 5.22 | 0.000* | ||||
Legend: *Significance level: α = 0.05
The heterogeneity statistics provide further insights into the variability of these effects. The Q-value of 310.62 with 20 degrees of freedom (df) indicates significant heterogeneity among the included studies (p < 0.0001), while the I² statistic of 93.56% points to a substantial proportion of variability that cannot be attributed to random sampling error alone (Viechtbauer & Cheung, 2010). This high level of heterogeneity suggests that the impact of scientific problem-solving approaches on student creativity varies considerably across different contexts, including educational settings, participant characteristics, and intervention designs.
Despite this variability, the consistently large effect sizes observed in both fixed and random-effects models underscore the overall effectiveness of scientific problem-solving approaches in fostering scientific creativity. This reinforces the value of such approaches in promoting innovative thinking and problem-solving skills, making them critical components of contemporary science education.
Subgroup analysis
Table 4 presents the effect sizes and corresponding forest plot derived from subgroup analyses of different scientific problem-solving approaches and STEM contexts. The overall effect size strongly favors experimental groups, with the estimate points consistently positioned to the right of the origin. This pattern underscores the effectiveness of these approaches in enhancing students’ scientific creativity and problem-solving skills, reflecting not only statistical significance but also meaningful practical impact.
Table 4. Subgroup analysis and forest plot of scientific problem-solving approaches
Authors | Hedges’ g | 95% CI | Forest Plot | |
|---|---|---|---|---|
Lower | Upper | |||
Chumo (2014) | 0.01 | -0.40 | 0.42 | |
Gunawan et al. (2018a) | 0.82 | 0.53 | 1.12 | |
Yang et al. (2016) | 0.43 | -0.22 | 1.07 | |
IBL | 0.43 | -0.62 | 1.48 | |
Dyson et al. (2016) | 2.49 | 1.63 | 3.35 | |
García and Fernández (2018) | 0.49 | 0.21 | 0.77 | |
Gunawan et al. (2018b) | 0.71 | 0.19 | 1.23 | |
Hu et al. (2013) | 2.11 | 1.63 | 2.59 | |
Siew and Ambo (2020) | 1.59 | 1.30 | 1.88 | |
Siew et al. (2017) | 1.45 | 1.08 | 1.82 | |
Combination | 1.43 | 0.64 | 2.22 | |
Demirhan and Sahin (2021) | 2.14 | 1.36 | 2.92 | |
Siew et al. (2017) | 0.94 | 0.59 | 1.29 | |
PBL | 1.49 | -6.11 | 9.09 | |
Aktamis and Ergi (2008) | 1.58 | 0.85 | 2.32 | |
Sari et al. (2019) | 5.02 | 4.12 | 5.93 | |
Siew and Ambo (2018) | 1.74 | 1.13 | 2.35 | |
Siew and Ambo (2020) | 0.18 | -0.07 | 0.44 | |
PjBL | 2.10 | -1.15 | 5.34 | |
Çalışıcı and Benzer (2021) | 1.04 | 0.39 | 1.68 | |
Doğan and Kahraman (2021) | 1.27 | 0.83 | 1.71 | |
Erol et al. (2022) | 1.94 | 1.34 | 2.53 | |
Kırıcı and Bakırcı (2021) | 3.20 | 2.44 | 3.96 | |
Ozkan and Topsakal (2019) | 0.07 | -0.40 | 0.53 | |
Tran et al. (2021) | -0.08 | -0.59 | 0.43 | |
STEM/STEAM | 1.21 | -0.06 | 2.49 | |
Combined effect size | 1.14 | 0.60 | 1.67 | |
Among the approaches, PjBL stands out with the largest effect size (g = 2.10), indicating its profound influence on fostering scientific creativity. This substantial effect can be attributed to the inherently student-centered nature of PjBL, which engages learners in sustained, authentic inquiry through complex project work. This approach requires students to apply scientific concepts in real-world contexts, mirroring the creative problem-solving tasks typical of professional scientific work (Aktamis & Ergi, 2008; Siew & Ambo, 2020). PjBL’s focus on open-ended exploration, iterative design, and collaborative teamwork encourages the development of higher-order cognitive skills, including planning, critical analysis, and creative production (Holm, 2011; Sari et al., 2019). These rich, contextualized learning experiences likely account for the robust impact observed, reinforcing the effectiveness of PjBL in nurturing innovative scientific thinkers.
Closely following, PBL demonstrates a strong effect size (g = 1.49). While PjBL emphasizes extended project work, PBL is structured around complex, ill-defined problems that require students to synthesize interdisciplinary knowledge, foster critical thinking, and develop innovative solutions (Akcay, 2009; Funa & Prudente, 2021). This approach immerses students in real-world scientific challenges, promoting metacognitive development and self-directed learning, which are critical for scientific creativity (Akınoğlu & Tandoğan, 2007; Siew et al., 2017). PBL’s emphasis on collaborative inquiry and reflective practice provides a fertile environment for creative problem-solving, aligning well with contemporary STEM education goals (Funa et al., 2024).
Interestingly, studies examining combined scientific problem-solving approaches (e.g., integrating PjBL, PBL, IBL, and STEM/STEAM) reveal a significant synergistic effect (g = 1.43). This finding highlights the potential for hybrid pedagogies to leverage the unique strengths of each method, creating multi-dimensional learning experiences that support creative thinking across diverse scientific contexts (Guo et al., 2020; Demirhan & Sahin, 2021). For instance, the project-centeredness of PjBL, the critical inquiry focus of PBL, the curiosity-driven exploration of IBL, and the interdisciplinary breadth of STEM/STEAM collectively provide students with a comprehensive problem-solving framework. This combination not only blends practical innovation with deep conceptual understanding but also prepares learners for complex, real-world scientific challenges (Hughes et al., 2022).
The STEM/STEAM approach alone also demonstrates a large effect (g = 1.21), reinforcing the value of interdisciplinary integration in nurturing creative problem-solving. STEM/STEAM education encourages students to blend scientific knowledge with engineering, technology, and artistic perspectives, fostering innovative thinking through hands-on, inquiry-driven learning (Dogan & Kahraman, 2021). This aligns with Henriksen’s (2014) assertion that STEM methods are particularly effective for cultivating creativity, as they provide students with diverse problem-solving frameworks that bridge multiple disciplines. This interdisciplinary approach enhances students’ capacity to generate innovative solutions by integrating knowledge from various scientific fields, aligning with real-world technological problem-solving (Hughes et al., 2022; Thuneberg et al., 2018).
IBL, on the other hand, while demonstrating a more moderate effect size (g = 0.43), remains a valuable approach for fostering scientific creativity, particularly in early educational contexts. Despite its comparatively lower impact, IBL emphasizes student-driven exploration and hypothesis testing, which can ignite curiosity and creative thinking when adequately supported by instructors (Gunawan et al., 2018a; Yang et al., 2016). However, the effectiveness of IBL can vary significantly depending on the level of instructional support, student readiness for self-directed inquiry, and the complexity of the problems presented (Chumo, 2014). This variability underscores the importance of carefully scaffolding IBL experiences to maximize their creative potential (Kim et al., 2023; Antonio & Prudente, 2024).
Building on the overall findings, a deeper examination of the sources of heterogeneity among the included studies (Table 5) reveals several critical contextual factors influencing the effectiveness of scientific problem-solving approaches in fostering scientific creativity. Key moderators include approaches, educational level, science domain, intervention duration, and learning format. Notably, studies conducted at the college level report the highest effect size (g = 3.21), suggesting that older students are particularly well-suited to scientific problem-solving interventions. This likely reflects their advanced cognitive capabilities, metacognitive awareness, and prior disciplinary knowledge, which together enhance their ability to navigate complex, open-ended scientific problems (Doğan & Kahraman, 2021). College students typically possess more developed critical thinking skills and a greater capacity for self-regulated learning, which enables them to engage effectively with the higher-order cognitive processes required for scientific creativity. In contrast, primary and secondary level studies show more moderate effects, possibly due to developmental differences that affect students’ readiness for self-directed, inquiry-based learning. Younger learners may struggle with the higher levels of abstraction and sustained problem-solving required in scientific problem-solving contexts, aligning with research that highlights the importance of cognitive maturity in creative thinking (Ozkan & Topsakal, 2019). This variation underscores the need for age-appropriate scaffolding when implementing scientific problem-solving at different educational levels.
Table 5. Moderator and subgroup analyses of scientific problem-solving approaches
Moderator | Subgroup | k | Test for Effect | Test for Heterogeneity | |||||
|---|---|---|---|---|---|---|---|---|---|
Hedges’ g | Std error | Variance | 95% CI | Q-value | Df(Q) | p-value | |||
Approaches | IBL | 3 | 0.43 | 0.24 | 0.19 | -0.62, 1.48 | 9.28 | 4 | 0.000** |
PjBL | 4 | 2.1 | 1.02 | 3.19 | -1.15, 5.34 | ||||
PBL | 2 | 1.49 | 0.6 | 0.63 | -6.11, 9.09 | ||||
STEM/STEM | 6 | 1.21 | 0.5 | 1.11 | -0.06, 2.49 | ||||
Combination* | 6 | 1.43 | 0.31 | 0.44 | 0.64, 2.22 | ||||
Educational Level | Primary | 8 | 1.22 | 0.23 | 0.45 | 0.68, 1.75 | 5.63 | 2 | 0.000** |
Secondary | 10 | 0.93 | 0.32 | 0.56 | 0.22, 1.65 | ||||
College | 3 | 3.21 | 0.91 | 2.14 | -0.69, 7.11 | ||||
Science Domain | Biology | 3 | 1.03 | 0.62 | 1.06 | -1.62, 3.68 | 0.27 | 3 | 0.000** |
Physics | 4 | 1.32 | 0.4 | 0.35 | 0.06, 2.58 | ||||
Science | 5 | 1.09 | 0.31 | 0.38 | 0.22, 1.95 | ||||
STEM/STEAM | 8 | 1.21 | 0.39 | 0.87 | 0.28, 2.14 | ||||
Duration of Intervention | 0–6 weeks | 9 | 1.85 | 0.5 | 1.02 | 0.70, 2.99 | 6.53 | 3 | 0.000** |
7–12 weeks | 8 | 1.05 | 0.26 | 0.52 | 0.43, 1.68 | ||||
More than 13 weeks | 2 | 1.28 | 0.84 | 1.33 | -9.41, 11.96 | ||||
Not reported | 2 | 0.35 | 0.35 | 0.19 | -4.11, 4.80 | ||||
Learning Format | Collaborative | 14 | 1.51 | 0.35 | 0.94 | 0.74, 2.27 | 2.32 | 2 | 0.000** |
Individual | 3 | 1.32 | 0.7 | 1.5 | -1.71, 4.35 | ||||
Combination* | 4 | 0.83 | 0.28 | 0.21 | -0.05, 1.72 | ||||
Note. *A combination of TPS pedagogical approaches (i.e., PjBL and STEM) or learning format (i.e., collaborative and individual activities); **Significant at p < 0.05
The impact of the science domain on the effectiveness of scientific problem-solving further highlights the importance of context. Studies situated within STEM/STEAM environments tend to show larger effect sizes compared to those focused on specific scientific disciplines like biology. This difference may stem from the inherently interdisciplinary nature of STEM/STEAM, which supports creative problem-solving by encouraging students to integrate diverse scientific concepts, technologies, and design processes in innovative ways (Gunawan et al., 2018b). STEM/STEAM curricula emphasize real-world problem-solving and cross-disciplinary thinking, providing an ideal environment for creativity by linking abstract scientific theories to practical, hands-on applications. In contrast, discipline-specific areas such as biology, which often emphasize conceptual understanding and factual knowledge over practical, technology-driven problem-solving, may limit opportunities for creative exploration. This suggests that scientific problem-solving approaches are particularly effective when aligned with curricula that prioritize technological design, systems thinking, and interdisciplinary applications, reinforcing the need for integrated learning frameworks that mirror authentic scientific and engineering practices.
Intervention duration also plays a critical role in determining the effectiveness of scientific problem-solving. Short-term interventions lasting up to six weeks consistently yield the highest effect sizes (g = 1.85), indicating that brief but intensive scientific problem-solving experiences can rapidly ignite creative thinking by immersing students in focused, high-impact learning environments (Yang et al., 2016). Short, time-bound projects often enhance students’ ability to generate innovative solutions, possibly due to the sense of urgency and concentrated effort such formats encourage. However, the effectiveness of these interventions appears to diminish over longer durations, suggesting potential cognitive fatigue, reduced novelty, or disengagement as students become accustomed to the learning format. This trend highlights the importance of pacing and dynamic instructional strategies to sustain creative engagement over extended periods.
Finally, the social dimension of learning emerges as a crucial moderator of the effectiveness of scientific problem-solving. Collaborative learning formats consistently outperform individual approaches, reflecting the essential role of social interaction, peer feedback, and collective problem-solving in driving technological creativity (Erol et al., 2022; Demirhan & Sahin, 2021). Group settings provide rich cognitive scaffolding that allows students to exchange diverse perspectives, test ideas, and co-create innovative solutions, all critical components of scientific problem-solving. This insight aligns with constructivist and sociocultural learning theories, which emphasize the transformative power of social discourse in developing higher-order thinking and creative skills (Vygotsky, 1978; Barrett et al., 2021). The positive impact of collaborative learning on scientific problem-solving underscores the importance of designing interventions that prioritize teamwork and real-world collaboration, reflecting the collaborative nature of modern scientific and technological innovation.
Publication bias analysis
Publication bias was assessed using both visual and statistical methods to ensure a comprehensive evaluation. The funnel plot (Fig. 2) revealed asymmetry, with an uneven distribution around the average effect size. This visual indication of bias was supported by a positive Kendall’s tau value of 0.41, indicating a moderate positive correlation between study size and effect size. The statistically significant p-value of 0.01 further suggests a relationship where larger studies report larger effect sizes, potentially reflecting some level of publication bias (Sterne et al., 2011).
Egger’s Regression Test provided additional evidence of funnel plot asymmetry, with a significant intercept of 5.99 (p = 0.00). This result suggests that smaller studies may have reported larger effect sizes than expected, further indicating potential publication bias. Similarly, the Begg-Mazumdar test revealed a moderate correlation consistent with the other findings.
Despite these indications of publication bias, the robustness of the findings was supported by Failsafe N values. Rosenthal’s and Orwin’s Failsafe N estimates (362 and 539, respectively) indicate that a substantial number of null-result studies would be required to significantly alter the overall effect size. Fisher’s Failsafe N value of 146, with a p-value of 0.00, further reinforces the stability of the observed effect sizes. These results suggest that, while some level of bias is present, the findings remain robust and relatively stable (Lin & Chu, 2017).
[See PDF for image]
Fig. 2
Standard error funnel plot and publication bias analysis
The presence of publication bias presents significant limitations on the generalizability of findings. For instance, asymmetry in the funnel plot indicates a potential underrepresentation of smaller studies, particularly those yielding non-significant or negative results, which may lead to an overestimation of the overall effect size. Such bias necessitates caution in extrapolating these results to broader educational contexts. To address these gaps, it is recommended that future research include unpublished data, pre-registered studies, and gray literature to minimize the impact of selective reporting. Additionally, replication studies in diverse educational settings and populations are crucial for validating the generalizability of the observed effects and mitigating the effects of publication bias. Incorporating mixed-method approaches in subsequent studies can enhance understanding of the factors underlying observed heterogeneities. Integrating qualitative evaluations with quantitative analyses could provide a more holistic perspective on the data, enriching the robustness and applicability of research outcomes.
Synthesis of scientific problem-solving approaches
This section synthesizes the dimensions of scientific problem-solving approaches and STEM contexts, focusing on their role in fostering scientific creativity. As presented in Table 6, the synthesis is structured around the core stages of the scientific method, in particular: problem definition, solution design, implementation, and evaluation, which collectively form a comprehensive framework for understanding how various educational strategies support creative thinking in science education (Sun et al., 2021). Examining these approaches highlights their unique contributions to developing scientific creativity, offering insights into practical strengths and areas for further enhancement.
Table 6. Dimensions and characteristics of scientific problem-solving approaches and STEM contexts
Approach | Authors, Year | Hedges’ g | Dimension | Characteristics |
|---|---|---|---|---|
PjBL | Aktamis and Ergi (2008), Sari et al. (2019), Siew and Ambo (2018, 2020) | 2.10 (Large) | Problem definition | Identify real-world, complex, and open-ended problems that are authentic to students’ contexts. |
Solution design | Develop creative, interdisciplinary solutions through iterative design, prototyping, and critical thinking. | |||
Implementation | Build and test solutions, adapt designs based on ongoing evaluation, and refine prototypes through multiple iterations. | |||
Evaluation | Assess outcomes, reflect on the learning process, and critically evaluate the effectiveness and impact of the final product. | |||
PBL | Demirhan and Sahin (2021) Siew et al. (2017) | 1.49 (Large) | Problem definition | Define complex problems, often related to real-world scenarios, encouraging collaborative inquiry and critical questioning. |
Solution design | Brainstorm potential solutions, develop conceptual models, and synthesize multiple viewpoints. | |||
Implementation | Engage in active, hands-on problem-solving, including simulations, role-playing, and real-world experimentation. | |||
Evaluation | Conduct reflective analysis, peer review, and self-assessment to validate the proposed solutions. | |||
Combination | Dyson et al. (2016), García and Fernández (2018), Gunawan et al. 2018b); Hu et al. (2013) | 1.43 (Large) | Problem definition | Integrate diverse perspectives to define multifaceted, interdisciplinary problems. |
Solution design | Use a mix of strategies (e.g., role-playing, everyday life activities) to generate innovative, context-rich solutions. | |||
Implementation | Implement diverse methods, including physical prototypes, digital simulations, and real-world experiments. | |||
Evaluation | Evaluate solutions through comprehensive assessments, including technical feasibility, creativity, and real-world applicability. | |||
STEM/STEAM | Çalışıcı and Benzer (2021), Doğan and Kahraman (2021), Ozkan and Topsakal (2019), Tran et al. (2021) | 1.21 (Large) | Problem definition | Identify interdisciplinary challenges requiring integration of scientific, technological, engineering, artistic, and mathematical knowledge. |
Solution design | Apply interdisciplinary insights to develop innovative solutions, including technical prototypes and artistic expressions. | |||
Implementation | Build functional prototypes, conduct complex experiments, and iteratively refine designs based on feedback. | |||
Evaluation | Assess the functionality, practicality, and impact of solutions within interdisciplinary contexts. | |||
IBL | Chumo (2014); Gunawan et al. 2018a); Yang et al. (2016) | 0.43 (Medium) | Problem definition | Formulate research questions based on observations, identify knowledge gaps, and define testable scientific problems. |
Solution design | Design controlled investigations, develop testable hypotheses, and plan systematic data collection. | |||
Implementation | Conduct experiments, gather and analyze data, and refine hypotheses based on empirical evidence. | |||
Evaluation | Draw conclusions, assess data quality, and communicate findings through scientific presentations or reports. |
PjBL emerges as particularly effective in promoting scientific creativity, demonstrated by its large effect size (g = 2.10). This approach emphasizes deep, authentic learning through real-world, student-driven projects that involve all stages of the problem-solving framework. For instance, Aktamis and Ergi (2008) demonstrated PjBL’s effectiveness by involving students in open-ended experiments on work-energy and simple machines, prompting them to identify real-world problems, design experimental solutions, and iteratively test their outcomes. Sari et al. (2019) further highlighted PjBL’s creative potential by guiding students to develop biochemistry projects using locally sourced materials, which required navigating complex, real-world constraints. This holistic approach aligns closely with the scientific problem-solving framework, as students not only define problems and generate solutions but also refine their designs based on iterative testing and feedback. This comprehensive engagement makes PjBL a particularly powerful method for fostering innovative thinking and practical application within STEM contexts.
PBL, with an effect size of g = 1.49, also aligns with this framework, although it emphasizes collaborative inquiry and problem identification rather than technical design. For example, Demirhan and Sahin (2021) documented PBL’s effectiveness in pre-service science teacher education, where students engaged in hands-on modeling of complex biological systems, such as the human circulatory and respiratory systems. This approach encourages students to collaboratively define problems, brainstorm solutions, and implement models that accurately represent real-world phenomena. Similarly, early childhood studies indicate that young learners can develop scientific creativity through unstructured, collaborative PBL tasks, fostering open-ended exploration and critical thinking. This approach’s focus on group discussion, active problem engagement, and iterative testing supports the key stages of scientific problem-solving, reinforcing its role in preparing learners to address complex, real-world challenges.
Combined approaches demonstrate a robust effect size (g = 1.43), reflecting the versatility of integrating multiple methods to enhance scientific creativity. This combination strategy leverages the strengths of different models to create a flexible and adaptive problem-solving framework. For instance, Dyson et al. (2016) utilized tabletop role-playing games to rapidly stimulate creative outputs, while Hu et al. (2013) incorporated everyday life activities to ground scientific concepts in familiar contexts. These hybrid methods effectively support all problem-solving stages, encouraging students to engage with diverse learning experiences across multiple disciplines. This integrative approach broadens the scope of possible solutions and enhances students’ ability to adapt to different problem contexts, making it a powerful strategy for fostering creative problem-solving skills in STEM education.
STEM/STEAM Approaches, with an effect size of g = 1.21, emphasize interdisciplinary problem-solving, requiring students to synthesize knowledge from science, technology, engineering, arts, and mathematics. This interdisciplinary focus is particularly effective in fostering scientific creativity, as it encourages students to apply their understanding across multiple domains. For example, Çalışıcı and Benzer (2021) demonstrated the impact of STEM/STEAM through activities requiring students to create technical products, such as telescopes and air filters, integrating engineering principles with scientific inquiry. Similarly, Doğan and Kahraman (2021) described how students developed functional prototypes, like thermos designs and mechanical clocks, combining scientific knowledge with practical engineering skills. These projects align closely with the problem-solving framework by emphasizing real-world applications, iterative design, and creative problem-solving, highlighting the value of STEM/STEAM in promoting innovative thinking within technology-rich educational settings.
IBL, represented by studies like Chumo (2014); Gunawan et al. (2018a), shows a moderate effect size (g = 0.43), reflecting its structured yet often less collaborative nature. This approach mirrors the scientific method, guiding students through question formulation, hypothesis testing, data collection, and analysis. For example, Chumo (2014) emphasized direct experimentation, where students formulated their own research questions and designed experiments to test biological theories. Similarly, Gunawan et al. (2018a) demonstrated the effectiveness of IBL in physics through virtual laboratory simulations, where students visualized complex scientific phenomena. Although IBL effectively develops critical thinking and scientific inquiry skills, its structured nature may limit creative exploration compared to more open-ended models like PjBL or STEM/STEAM. Nonetheless, IBL remains a crucial component of the scientific problem-solving framework, offering foundational skills necessary for rigorous scientific investigation.
Together, these approaches illustrate the diverse pathways through which scientific problem-solving can be effectively implemented in science education. Each method offers distinct strengths, from the deep, context-rich projects of PjBL to the structured scientific inquiry of IBL, highlighting the importance of selecting the appropriate pedagogical model based on the desired educational outcomes and the specific needs of learners. By aligning these approaches with the stages of the scientific problem-solving framework, educators can design and implement interventions that cultivate scientific creativity and innovative thinking within STEM contexts.
Implications for practice and further research
The findings of this meta-analysis underscore the importance of integrating scientific problem-solving approaches, including PjBL, PBL, STEM/STEAM, and IBL, in fostering scientific creativity across diverse educational contexts. For instance, high school science teachers could leverage PjBL by challenging students to design eco-friendly waste management systems, thereby integrating real-world applications with collaborative problem-solving. At the elementary level, IBL can be effectively used to encourage younger students to investigate natural phenomena, such as the impact of sunlight on plant growth, nurturing early curiosity and scientific reasoning. These approaches provide flexible, context-specific strategies for cultivating creative thinking, allowing educators to adapt their teaching methods based on the unique needs and interests of their students. For example, a robotics instructor might integrate STEM/STEAM activities to promote interdisciplinary thinking, while a chemistry teacher could use PBL to explore real-world challenges like water purification or environmental sustainability. These practices not only enhance students’ problem-solving skills but also prepare them to address complex scientific and technological challenges in their future careers.
To maximize the impact of these approaches, educators should consider the importance of scaffolding, real-world relevance, and interdisciplinary connections in their lesson designs. Providing professional development opportunities, such as workshops on integrating STEM/STEAM principles or inquiry-based techniques, can further support teachers in implementing these strategies effectively. Additionally, educational leaders should prioritize curricula that emphasize scientific creativity and problem-solving, aligning instructional practices with the demands of modern scientific and technological fields.
Future research should explore the long-term impacts of these approaches on students’ scientific creativity and problem-solving abilities across different grade levels and disciplines. For instance, longitudinal studies could provide insights into how sustained exposure to these methods influences creativity development over time. Comparative studies are also needed to assess the differential effectiveness of these approaches in various cultural and educational settings, such as urban versus rural schools, to better understand the role of context in shaping learning outcomes.
To address the heterogeneity in effect sizes observed in this meta-analysis, future studies should adopt mixed-method designs. Combining quantitative meta-analytic techniques with qualitative case studies can reveal deeper insights into the contextual factors that influence the effectiveness of scientific problem-solving approaches. For example, mixed-method studies could examine how individual components of these frameworks, such as collaboration or inquiry, contribute to scientific creativity in different educational contexts. Meta-regression analyses could further help identify key moderators, such as the duration of interventions, teacher experience, or student demographics, that explain variability in effect sizes.
Moreover, researchers should work toward developing more nuanced frameworks that integrate elements of PjBL, PBL, STEM/STEAM, and IBL, reflecting the complex, interdisciplinary nature of real-world scientific problem-solving. For example, a hybrid framework might combine PBL’s focus on problem definition with STEM/STEAM’s emphasis on interdisciplinary connections to foster innovation in engineering design. Understanding the role of teacher training and professional development in the successful implementation of these approaches is also crucial. Equipping educators with the skills and knowledge needed to apply these strategies effectively can provide valuable insights into sustaining and scaling innovative pedagogical practices, ultimately supporting the development of creative, scientifically literate learners.
Limitations of the current review
This meta-analysis, while offering valuable insights into the effectiveness of scientific problem-solving approaches in fostering scientific creativity, has several limitations. First, the relatively small number of studies specifically focused on scientific creativity within these approaches limits the generalizability of the findings. The included studies vary significantly in terms of educational levels, instructional settings, research designs, and sample sizes, introducing heterogeneity that may affect the robustness of the estimated effect sizes. Additionally, the emphasis on experimental and quasi-experimental designs, while enhancing the internal validity of the findings, excludes rich qualitative insights that could provide a deeper understanding of how these approaches function in diverse educational contexts. This narrow focus overlooks critical aspects of student experiences, teacher perspectives, and the complex, context-dependent nature of scientific creativity.
Moreover, the absence of longitudinal studies restricts the ability to assess the sustained impact of these approaches on students’ creativity and problem-solving skills over time. Without long-term data, it remains unclear whether the observed gains in creativity persist beyond the immediate context of the interventions. Furthermore, the potential for publication bias, where studies with significant or favorable results are more likely to be published, may have inflated the overall effect sizes reported in this review. This bias underscores the need for more comprehensive, transparent reporting of both positive and null findings in future research.
To address these limitations, future studies should expand the scope of inquiry to include a broader range of methodologies, including qualitative research and mixed-method designs. For instance, qualitative case studies could provide insights into the lived experiences of students and teachers, capturing the nuanced dynamics of scientific creativity that are often missed in purely quantitative analyses. Such studies might explore how these approaches are implemented across different cultural and educational contexts, how teachers perceive their effectiveness and feasibility, and how students engage with the problem-solving processes. Additionally, integrating longitudinal designs would help clarify the long-term effects of these approaches on scientific creativity, offering a more comprehensive understanding of their impact.
Conclusion
This meta-analysis demonstrates the effectiveness of scientific problem-solving approaches, including PjBL, PBL, STEM/STEAM, and IBL, in fostering scientific creativity among students. These approaches vary in their impact, with PjBL and PBL generally yielding the most substantial effect sizes, reflecting their emphasis on real-world problem-solving, interdisciplinary thinking, and hands-on experimentation. Despite limitations such as the relatively small number of studies, varying research designs, and potential publication bias, the findings underscore the critical role of these pedagogical strategies in enhancing students’ creative capacities and problem-solving skills. This evidence supports the integration of scientific problem-solving approaches into science education curricula, providing a robust foundation for cultivating innovative thinking and practical scientific skills among learners. Future studies should further investigate these approaches across diverse educational contexts to fully capture their long-term impact and effectiveness.
Acknowledgements
The authors wish to extend their heartfelt appreciation to the Department of Science and Technology - Science Education Institute (DOST-SEI) through the Capacity Building Program in Science and Mathematics Education (CBPSME) for the invaluable support in this research. They also gratefully acknowledge the reviewers and editors of this journal for the insightful comments and contributions, which have greatly enriched the quality of this manuscript.
Author contributions
The contributions of the authors to this research are as follows: FIP, AAP, and RS were responsible for conceptualization. The reframing of objectives, research questions, and gap identification was carried out by FIP and DA. FIP, AAP, RS, and DA conducted the literature search procedures, coding, and analysis with supervision from LR and MP. The original draft was written by FIP and DA. Finally, AAP, RS, LR, and MP reviewed and edited the original draft. All authors have read and agreed to the published version of the manuscript.
Funding
Not applicable.
Data availability
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.
Declarations
Competing interests
The authors declare no competing interests.
Abbreviations
Inquiry Based Learning
Problem-Based Learning
Project-Based Learning
Science, Technology, Engineering, Mathematics
Science, Technology, Engineering, Arts, Mathematics
Quasi-experimental Research
Mixed-methods Research
Preferred Reporting Items for Systematic Reviews and Meta-Analyses
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
Abulibdeh, A; Zaidan, E; Abulibdeh, R. Navigating the confluence of artificial intelligence and education for sustainable development in the era of industry 4.0: Challenges, opportunities, and ethical dimensions. Journal of Cleaner Production; 2024; 437, 140527. [DOI: https://dx.doi.org/10.1016/j.jclepro.2023.140527]
Acut, D. Developing SIPCaR projects utilizing modern technologies: Its impact to students’ engagement, R&D skills, and learning outcomes. LUMAT: International Journal on Math Science, and Technology Education; 2022; 10,
Acut, D; Antonio, R. Effectiveness of Science-Technology-Society (STS) approach on students’ learning outcomes in science education: Evidence from a meta-analysis. Journal of Technology and Science Education; 2023; 13,
Akçay, B. Problem-Based learning in science education. Journal of Turkish Science Education; 2009; 6,
Akınoğlu, O; Tandoğan, RÖ. The effects of problem-based active learning in science education on students’ academic achievement, attitude and concept learning. Eurasia Journal of Mathematics, Science and Technology Education; 2007; 3,
Aktamis, H., & Ergin, O. (2008). The effect of scientific process skills education on students’ scientific creativity, science attitudes and academic achievements. Asia-Pacific Forum on Science Learning and Teaching, 9(1). https://www.eduhk.hk/apfslt/v9_issue1/aktamis/index.htm#abstract
Ammar, M; Al-Thani, NJ; Ahmad, Z. Role of pedagogical approaches in fostering innovation among K-12 students in STEM education. Social Sciences & Humanities Open; 2024; 9, 100839. [DOI: https://dx.doi.org/10.1016/j.ssaho.2024.100839]
Antonio, RP; Prudente, MS. Effects of inquiry-based approaches on students’ higher-order thinking skills in science: A meta-analysis. International Journal of Education in Mathematics, Science and Technology; 2024; 12,
Armstrong, P. (2010). Bloom’s Taxonomy. Vanderbilt University Center for Teaching. https://cft.vanderbilt.edu/guides-sub-pages/blooms-taxonomy/
Barrett, M. S., Creech, A., & Zhukov, K. (2021). Creative collaboration and collaborative creativity: A systematic literature review. Frontiers in Psychology, 12. https://doi.org/10.3389/fpsyg.2021.713445
Begg, CB; Mazumdar, M. Operating characteristics of a rank correlation test for publication bias. Biometrics; 1994; 50,
Bi, H; Mi, S; Lu, S; Hu, X. Meta-analysis of interventions and their effectiveness in students’ scientific creativity. Thinking Skills and Creativity; 2020; 38, 100750. [DOI: https://dx.doi.org/10.1016/j.tsc.2020.100750]
Borenstein, M., Hedges, L., Higgins, J. P., & Rothstein, H. R. (2009). Introduction to meta-analysis. Wiley & Sons. https://doi.org/10.1002/9780470743386
Çalışıcı, S., & Benzer, S. (2021). The effects of STEM applications on the environmental attitudes of the 8th year students, scientific creativity and science achievements. Malaysian Online Journal of Educational Sciences, 9(1). https://jml.um.edu.my/index.php/MOJES/article/view/28217
Cheng, M. M. H. (2022). An overview of STEM education in Asia. In M. M. H. Cheng, C. Buntting, & A. Jones (Eds.), Concepts and practices of STEM education in Asia. Springer. https://doi.org/10.1007/978-981-19-2596-2_1
Chumo, C. C. (2014). Effects of practical investigation on scientific creativity amongst secondary schools biology students in Kericho district, Kenya. Journal of Education and Practice, 5(8). https://www.iiste.org/Journals/index.php/JEP/article/view/11639
Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Lawrence Erlbaum.
Cooper, H. M. (2010). Research synthesis and meta-analysis: A step-by-step approach. Sage.
Demirhan, E; Şahin, F. The effects of different kinds of hands-on modeling activities on the academic achievement, problem-solving skills, and scientific creativity of prospective science teachers. Research in Science Education; 2021; 51,
Dogan, A; Kahraman, E. The effect of STEM activities on the scientific creativity of middle school students. International Journal of Curriculum and Instruction; 2021; 13,
Dyson, SB; Chang, YL; Chen, HC; Hsiung, HY; Tseng, CC; Chang, JH. The effect of tabletop role-playing games on the creative potential and emotional creativity of Taiwanese college students. Thinking Skills and Creativity; 2016; 19, pp. 88-96. [DOI: https://dx.doi.org/10.1016/j.tsc.2015.10.004]
Egger, M; Smith, GD; Schneider, M; Minder, C. Bias in meta-analysis detected by a simple, graphical test. British Medical Journal; 1997; 315,
Erol, A; Erol, M; Başaran, M. The effect of STEAM education with Tales on problem solving and creativity skills. European Early Childhood Education Research Journal; 2022; 31,
Fernandez, S. M., Madelo, P. K., Lu Suico, R. A., Cane, J. F., Magsayo, J., Capuyan, M., Siew, N. M., & Acut, D. (2024). Thinking what no one else has thought: Investigating the scientific creativity of primary school students in a science class. Center for Educational Policy Studies Journal. https://doi.org/10.26529/cepsj.1514
Fisher, R. A. (1932). Statistical methods for research workers (4th ed.). Oliver & Boyd.
Funa, AA; Prudente, MS. Effectiveness of problem-based learning on secondary students’ achievement in science: A meta-analysis. International Journal of Instruction; 2021; 14,
Funa, A. A., Roleda, L. S., & Prudente, M. S. (2024). Integrated science, technology, engineering, and mathematics—problem-based learning—education for sustainable development (I-STEM-PBL-ESD) framework. In Y. S. Ong, T. T. M. Tan, & Y. J. Lee (Eds.), A diversity of pathways through science education. Springer. https://doi.org/10.1007/978-981-97-2607-3_9
García, PC; Fernandez, TG. Makerspaces and scientific creativity level of middle school students. Global Journal of Arts Education; 2018; 8,
Gopnik, A. Scientific thinking in young children: Theoretical advances, empirical research, and policy implications. Science; 2012; 337,
Gunawan, G., Nisrina, N., & Suranti, N. (2018a). Enhancing students’ creativity in physics classroom using virtual laboratory. Advances in Social Science, Education and Humanities Research. https://doi.org/10.2991/ictte-18.2018.67
Gunawan, N., Harjono, A., Sahidu, H., & Nisrina, N. (2018b). Improving students’ creativity using cooperative learning with virtual media on static fluida concept. Journal of Physics Conference Series, 1006, 012016. https://doi.org/10.1088/1742-6596/1006/1/012016
Guo, P; Saab, N; Post, LS; Admiraal, W. A review of project-based learning in higher education: Student outcomes and measures. International Journal of Educational Research; 2020; 102, 101586. [DOI: https://dx.doi.org/10.1016/j.ijer.2020.101586]
Hedges, L., & Olkin, I. (1985). Statistical methods for meta-analysis. Academic Press.
Henriksen, D. Full STEAM ahead: Creativity in excellent STEM teaching practices. STEAM; 2014; 1,
Hernández-Torrano, D; Ibrayeva, L. Creativity and education: A bibliometric mapping of the research literature (1975–2019). Thinking Skills and Creativity; 2020; 35, 100625. [DOI: https://dx.doi.org/10.1016/j.tsc.2019.100625]
Hmelo-Silver, CE. Problem-Based learning: What and how do students learn??. Educational Psychology Review; 2004; 16,
Holm, M. Project-based instruction: A review of the literature on effectiveness in pre-kindergarten through = 12th grade classrooms. InSight: Rivier Academic Journal; 2011; 7,
Hu, W; Adey, P. A scientific creativity test for secondary school students. International Journal of Science Education; 2002; 24,
Hu, W; Wu, B; Jia, X; Yi, X; Duan, C; Meyer, W; Kaufman, JC. Increasing students’ scientific creativity: The learn to think intervention program. The Journal of Creative Behavior; 2013; 47,
Hughes, B. S., Corrigan, M. W., Grove, D., Andersen, S. B., & Wong, J. T. (2022). Integrating arts with STEM and leading with STEAM to increase science learning with equity for emerging bilingual learners in the united States. International Journal of STEM Education, 9(1). https://doi.org/10.1186/s40594-022-00375-7
Kaufman, J. C., & Sternberg, R. J. (Eds.). (2019). The Cambridge handbook of creativity (2nd ed.). Cambridge University Press. https://doi.org/10.1017/9781316979839
Kim, S; Raza, M; Seidman, E. Improving 21st-century teaching skills: The key to effective 21st-century learners. Research in Comparative and International Education; 2019; 14,
Kim, J; Im, H; Ahn, D; Cho, S.
Kırıcı, M. G., & Bakırcı, H. (2021). The effect of STEM supported research-inquiry-based learning approach on the scientific creativity of 7th grade students. Journal of Pedagogical Research, 1–17. https://doi.org/10.33902/jpr.2021067921
Larmer, J; Mergendoller, JR. Seven essentials for project-based learning. Educational Leadership; 2010; 68,
Lin, L; Aloe, AM. Evaluation of various estimators for standardized mean difference in meta-analysis. Statistics in Medicine; 2020; 40,
Lin, L; Chu, H. Quantifying publication bias in meta-analysis. Biometrics; 2017; 74,
Long, H; Kerr, BA; Emler, TE; Birdnow, M. A critical review of assessments of creativity in education. Review of Research in Education; 2022; 46,
Moher, D; Liberati, A; Tetzlaff, J; Altman, DG. Preferred reporting items for systematic reviews and meta-analyses: The PRISMA statement. British Medical Journal; 2009; 339, pp. 1-8. [DOI: https://dx.doi.org/10.1136/bmj.b2535]
Orwin, RG. A fail-safe N for effect size in meta-analysis. Journal of Educational Statistics; 1983; 8,
Ozkan, G; Topsakal, UU. Exploring the effectiveness of STEAM design processes on middle school students’ creativity. International Journal of Technology and Design Education; 2019; 31,
P21’s Framework for 21st Century Learning (2019). https://www.battelleforkids.org/wp-content/uploads/2023/11/P21_Framework_Brief.pdf
Panergayo, AAE; Prudente, MS. Effectiveness of design-based learning in enhancing scientific creativity in STEM education: A meta-analysis. International Journal of Education in Mathematics Science and Technology; 2024; 12,
Panjaitan, MB; Siagian, A. The effectiveness of inquiry based learning model to improve science process skills and scientific creativity of junior high school students. Journal of Education and E-Learning Research; 2020; 7,
Popper, K. (1959). The logic of scientific discovery. Abingdon-on-thames, United Kingdom: Routledge. https://doi.org/10.1063/1.3060577
Rosenthal, R. The file drawer problem and tolerance for null results. Psychological Bulletin; 1979; 86,
Runco, M. A. (2007). Creativity: Theories and themes: Research, development, and practice. Elsevier Academic Press.
Sandri, OJ. Exploring the role and value of creativity in education for sustainability. Environmental Education Research; 2013; 19,
Sari, D. K., Ibrahim, A., & Wancik, K. A. (2019). Designing biochemistry project with local materials to increased student’s creativity. Journal of Physics Conference Series, 1166, 012007. https://doi.org/10.1088/1742-6596/1166/1/012007
Schmidt, F. L., & Hunter, J. E. (2015). Methods of meta-analysis: Correcting error and bias in research findings. https://doi.org/10.4135/9781483398105
Siew, NM; Ambo, N. Development and evaluation of an integrated project-based and STEM teaching and learning module on enhancing scientific creativity among fifth graders. Journal of Baltic Science Education; 2018; 17,
Siew, NM; Ambo, N. The scientific creativity of fifth graders in a STEM Project-Based cooperative learning approach. Problems of Education in the 21st Century; 2020; 78,
Siew, NM; Chin, MK; Sombuling, A. The effects of problem based learning with cooperative learning on preschoolers’ scientific creativity. Journal of Baltic Science Education; 2017; 16,
Simonton, DK. Taking the U.S. Patent office criteria seriously: A quantitative three-criterion creativity definition and its implications. Creativity Research Journal; 2012; 24,
Stehle, S. M., & Peters-Burton, E. E. (2019). Developing student 21st century skills in selected exemplary inclusive STEM high schools. International Journal of STEM Education, 6(1). https://doi.org/10.1186/s40594-019-0192-1
Sterne, JC; Sutton, AJ; Ioannidis, JPA; Terrin, N; Jones, DR; Lau, J; Carpenter, J; Rucker, G; Harbord, RM; Schmid, CH; Tetzlaff, J; Deeks, JJ; Peters, J; Macaskill, P; Schwarzer, G; Duval, S; Altman, DG; Moher, D; Higgins, JPT. Recommendations for examining and interpreting funnel plot asymmetry in meta-analyses of randomised controlled trials. British Medical Journal; 2011; 343, d4002. [DOI: https://dx.doi.org/10.1136/bmj.d4002]
Stolaki, A; Satratzemi, M; Xinogalos, S. Examining the effects of creativity, collaboration, creative diversity and autonomy on team creative performance. Thinking Skills and Creativity; 2023; 50, 101415. [DOI: https://dx.doi.org/10.1016/j.tsc.2023.101415]
Sun, M; Wang, M; Wegerif, R; Peng, J. How do students generate ideas together in scientific creativity tasks through computer-based Mind mapping?. Computers & Education; 2021; 176, 104359. [DOI: https://dx.doi.org/10.1016/j.compedu.2021.104359]
Suurmond, R; van Rhee, H; Hak, T. Introduction, comparison and validation of metaessentials: A free and simple tool for meta-analysis. Research Synthesis Methods; 2017; 8,
Thuneberg, H; Salmi, H; Bogner, F. How creativity, autonomy and visual reasoning contribute to cognitive learning in a STEAM hands-on inquiry-based math module. Thinking Skills and Creativity; 2018; 29, pp. 153-160. [DOI: https://dx.doi.org/10.1016/j.tsc.2018.07.003]
Torrance, E. P. (1990). Torrance tests of creative thinking. Scholastic Testing Service.
Tran, NH; Huang, CF; Hung, JF. Exploring the effectiveness of STEAM-Based courses on junior high school students’ scientific creativity. Frontiers in Education; 2021; [DOI: https://dx.doi.org/10.3389/feduc.2021.666792]
Tuyet, TLT; Thi, KN; Duc, HT; Giao, XTT; Thi, NV; Tuong, VHK; Nhu, UDT; Mai, ALT. STEM education in primary schools of Southeast Asian countries: An analysis of scientific publications in the scopus database from 2000 to 2022. Eurasia Journal of Mathematics Science and Technology Education; 2024; 20,
Van Rhee, H. J., Suurmond, R., & Hak, T. (2015). User manual for Meta-Essentials: Workbooks for meta-analysis (Version 1.4). Erasmus Research Institute of Management. www.erim.eur.nl/research-support/meta-essentials
Viechtbauer, W; Cheung, MW. Outlier and influence diagnostics for meta-analysis. Research Synthesis Methods; 2010; 1,
Vo, DV; Simmie, GM. Assessing scientific inquiry: A systematic literature review of tasks, tools and techniques. International Journal of Science and Mathematics Education; 2024; 23, pp. 871-906. [DOI: https://dx.doi.org/10.1007/s10763-024-10498-8]
Vygotsky, L. S. (1978). Mind in society: The development of higher psychological processes. Harvard University Press.
Ward, TB. Creative cognition as a window on creativity. Methods; 2007; 42,
Xu, S; Reiss, MJ; Lodge, W. Comprehensive scientific creativity assessment (C-SCA): A new approach for measuring scientific creativity in secondary school students. International Journal of Science and Mathematics Education; 2024; [DOI: https://dx.doi.org/10.1007/s10763-024-10469-z]
Yang, KK; Lee, L; Hong, ZR; Lin, HS. Investigation of effective strategies for developing creative science thinking. International Journal of Science Education; 2016; 38,
Yildiz, C; Yildiz, TG. Exploring the relationship between creative thinking and scientific process skills of preschool children. Thinking Skills and Creativity; 2021; 39, 100795. [DOI: https://dx.doi.org/10.1016/j.tsc.2021.100795]
Zhong, B., Liu, X., Xia, L., & Sun, W. (2022). A proposed taxonomy of teaching models in STEM education: Robotics as an example. SAGE Open, 12(2). https://doi.org/10.1177/21582440221099525
© The Author(s) 2025. This work is published under http://creativecommons.org/licenses/by/4.0/ (the “License”). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License.