You are currently viewing a new version of our website. To view the old version click .
MDPIMDPI
Search all articles
Journal of Intelligence
Download PDF
  • Systematic Review
  • Open Access

15 November 2025

Mapping the Scaffolding of Metacognition and Learning by AI Tools in STEM Classrooms: A Bibliometric–Systematic Review Approach (2005–2025)

,
,
and
Department of Mathematics, Natural Sciences and Technology Education, Faculty of Education, University of the Free State, Bloemfontein 9301, South Africa
*
Author to whom correspondence should be addressed.
J. Intell.2025, 13(11), 148;https://doi.org/10.3390/jintelligence13110148 
(registering DOI)
This article belongs to the Section Studies on Cognitive Processes
Version Notes
Order Reprints

Abstract

This study comprehensively analyses how AI tools scaffold and share metacognitive processes, thereby facilitating students’ learning in STEM classrooms through a mixed-method research synthesis combining bibliometric analysis and systematic review. Using a convergent parallel mixed-methods design, the study draws on 135 peer-reviewed articles published between 2005 and 2025 to map publication trends, author and journal productivity, keyword patterns, and theoretical frameworks. Data were retrieved from Scopus and Web of Science using structured Boolean searches and analysed using Biblioshiny and VOSviewer. Guided by PRISMA 2020 protocols, 24 studies were selected for in-depth qualitative review. Findings show that while most research remains grounded in human-centred conceptualisations of metacognition, there are emerging indications of posthumanist framings, where AI systems are positioned as co-regulators of learning. Tools like learning analytics, intelligent tutoring systems, and generative AI platforms have shifted the discourse from individual reflection to system-level regulation and distributed cognition. The study is anchored in Flavell’s theory of metacognition, General Systems Theory, and posthumanist perspectives to interpret this evolution. Educational implications highlight the need to reconceptualise pedagogical roles, integrate AI literacy in teacher preparation, and prioritise ethical, reflective AI design. The review provides a structured synthesis of theoretical, empirical, and conceptual trends, offering insights into how human–machine collaboration is reshaping learning by scaffolding and co-regulating students’ metacognitive development in STEM education.

1. Introduction

The 21st century has seen a profound transformation in education through adaptive teaching, digital technologies, and artificial intelligence (AI), especially in STEM (Science, Technology, Engineering, and Mathematics) subjects (; ). As STEM education shifts toward personalised and reflective learning, metacognition—learners’ ability to monitor, control, and direct their thinking—has become increasingly critical (). Building on ’s () foundational work, metacognition is understood to involve both knowledge of one’s cognitive processes and the regulation of strategies such as planning, monitoring, and evaluation ().
Metacognition is particularly important in STEM, where problem-solving and experimentation dominate over rote learning (; ; ). Numerous studies link metacognitive awareness to improved conceptual understanding, motivation, reasoning, and academic performance (; ; ; ). Interventions such as mentoring and structured reflection further enhance self-regulation and comprehension in technology-rich environments (). However, fostering metacognition remains challenging due to limited teacher training, time, and resources, as well as the difficulty of observing internal cognitive processes (; ). These challenges underscore the need for innovative approaches—such as AI-powered tools—to support metacognitive development in STEM classrooms.
Recent advancements in AI offer promising avenues to address these challenges. AI-powered tools, such as Intelligent Tutoring Systems (ITS), adaptive platforms, learning analytics dashboards, and conversational agents, can scaffold learners’ metacognitive development by providing personalised feedback, real-time monitoring, and strategic prompts (; ; , , ). For example, platforms such as Carnegie Learning and ALEKS promote both content mastery (object level) and reflective monitoring (meta level) (). Adaptive AI approaches, including deep reinforcement learning, have also shown potential to enhance metacognitive outcomes (). Bibliometric evidence further indicates a growing focus on ITS and related technologies that enable teachers to observe and support metacognitive behaviours ().
However, ethical concerns remain if AI prioritises automation and speed over reflection, prompting scholars to call for responsible, transparent, and learner-centred AI integration (; ; ). Given both the promise and pitfalls of AI in supporting metacognition, it is crucial to understand how this intersection has evolved. While numerous empirical studies have examined individual AI tools or specific metacognitive strategies, existing reviews tend to focus broadly on AI in education without explicitly addressing how AI supports metacognition within STEM classrooms (; ). However, these reviews rarely integrate bibliometric mapping with systematic qualitative synthesis, leaving gaps in understanding publication trends, theoretical framings, and the evolution of concepts over time. This study addresses these gaps by combining bibliometric analysis of 135 publications (2005–2025) with a systematic review of 24 core studies, offering a comprehensive two-decade overview of how AI tools scaffold and co-regulate metacognition in STEM education. By doing so, it provides valuable insights for researchers, educators, and policymakers seeking to design effective, ethically grounded AI-supported learning environments.

1.1. Theoretical Framework

This study is guided by three complementary frameworks: Flavell’s Theory of Metacognition (1979), General Systems Theory (GST), and Human-Centred vs. Posthumanist AI paradigms. Together, they provide a multidimensional lens for understanding how AI tools scaffold metacognition in STEM education.
Rather than redefining metacognition, we apply Flavell’s model (as introduced above) to interpret how AI tools support regulation at both the object level (task performance) and the meta level (strategic monitoring), following ’ () distinction (; ). ’ () distinction between the object level (task performance) and meta level (strategic regulation) is integrated here to illustrate how AI operates simultaneously in content mastery and reflective monitoring.
GST conceptualises STEM classrooms as interconnected, adaptive systems where students, teachers, technologies, and pedagogies interact dynamically (, ; ). This perspective helps map how AI tools interact with cognitive and instructional subsystems, creating system-level effects that can be traced across the literature.
Ultimately, the human-centred paradigm regards AI as a means of supporting human agency under transparent and ethical oversight (; ). In contrast, the posthumanist paradigm challenges human–machine binaries by conceptualising learning as a co-regulated process distributed between humans and AI systems (; ). We use these paradigms to interpret whether AI is framed as a supportive tool (human-centred) or a co-agent in reflective processes (posthumanist).

1.2. Literature Review

This section reviews key studies on metacognition in STEM education and the role of AI in supporting metacognitive processes. It highlights evidence of improved learning outcomes through the use of metacognitive strategies. It examines how AI tools have been integrated to enhance self-regulation, reflection, and critical thinking in STEM contexts.

1.2.1. Metacognition and STEM Education

Metacognition—the ability to plan, monitor, and evaluate one’s thinking—is critical for problem-solving and conceptual understanding in STEM contexts (; ). Studies consistently show its positive impact on academic performance across disciplines, including mathematics, science, chemistry, and physics (; ; ; ; ).
Research also links metacognitive skills to motivation and reasoning (; ; ), highlighting benefits for low-achieving and underrepresented groups (; ; ). Despite this, challenges persist due to inconsistent measurement methods (), theoretical fragmentation (), and variability across contexts (; ). These gaps underscore the need for integrated theoretical approaches (), which align with the objectives of the present review.

1.2.2. Artificial Intelligence and Metacognition in STEM Education

AI systems—ranging from intelligent tutoring systems and adaptive platforms to generative AI tools like ChatGPT (versions GPT-4.1, GPT-4.5 and GPT-5)—are increasingly used to scaffold metacognition by providing real-time feedback and strategic prompts (; ). Learning analytics dashboards and conversational agents help externalise metacognitive processes by tracking learner progress and supporting reflection (; ; ).
Recent studies demonstrate AI’s role in enhancing self-regulation, motivation, and retention through metacognitive prompts, chatbot feedback, and ME-CoT approaches (; ; ). Furthermore, AI has been linked to computational and critical thinking skills, fostering deeper cognitive engagement (; ). However, concerns remain regarding cognitive overload, reduced learner autonomy, and algorithmic bias if AI prioritises automation over deep reflection (; ; ).
While several bibliometric reviews have examined AI in education broadly, none focus explicitly on the intersection of AI and metacognition in STEM education (; ). This study addresses that gap by mapping publication trends, theoretical frameworks, and conceptual shifts, with attention to the emerging discourse moving from human-centred approaches to posthumanist framings. The following research questions guide the review:
  • What are the trends in publication on metacognition in STEM education between 2005 and 2025?
  • What are the most frequently occurring AI-related concepts and tools in the literature on metacognition in STEM education?
  • Which journals, authors, and countries contribute most to the literature on AI and metacognition in STEM education?
  • Which theoretical frameworks are most frequently associated with studies on AI and metacognition in STEM, and how have these evolved?
  • How have keywords and conceptual language in the literature shifted from human-centred to posthumanist paradigms in the context of AI and metacognition?

2. Materials and Methods

This study adopted a convergent parallel mixed-method research synthesis design, combining bibliometric analysis with a systematic review. Both components were conducted independently but addressed complementary aspects of the research questions. The bibliometric analysis provided a quantitative map of trends, conceptual clusters, and publication networks spanning the period from 2005 to 2025. In parallel, the systematic review offered qualitative insights into how theoretical frameworks (including posthumanist paradigms) were employed in studies that integrated AI tools with metacognitive strategies in STEM classrooms. Findings from both strands were integrated during interpretation, allowing for cross-validation and thematic synthesis of patterns emerging from both data sources.

2.1. Data Sources and Search Strategy

Searches were conducted in Scopus and Web of Science, two multidisciplinary databases recognised for their credibility and extensive coverage of peer-reviewed research (). These databases were selected to ensure a comprehensive and diverse capture of studies at the intersection of AI, metacognition, and STEM education. Boolean operators and structured search strings were applied to retrieve highly relevant documents (see Supplementary Table S1 for full details).
This combined bibliometric–systematic review approach was chosen because it provides distinct advantages over traditional narrative or purely systematic reviews. Bibliometric analysis provides an objective, data-driven map of research trends, influential contributors, and conceptual networks, while systematic reviews offer deeper thematic and theoretical insights. Together, these methods enhance the rigour and validity of findings by triangulating quantitative mapping with qualitative synthesis, yielding a comprehensive understanding of how AI has been used to scaffold metacognition in STEM education.

2.2. Inclusion and Exclusion Criteria

To ensure relevance and rigor, documents were included based on the following criteria: (1) published between 2005 and 2025, (2) written in English, (3) peer-reviewed journal articles or conference proceedings, (4) focused on metacognitive development in STEM education or broader education and learning contexts of AI integration, and (5) explicitly referenced theoretical or conceptual frameworks. Documents were excluded if they were (a) not in English, (b) editorials, reviews, or book chapters, (c) not directly related to both AI and metacognition, or (d) lacked educational context. We selected 2005 as the starting point because this period marks the emergence and wider adoption of AI-based tools in education, including intelligent tutoring systems, adaptive learning platforms, and early learning analytics applications (e.g., ). Furthermore, bibliometric scoping revealed that publications addressing AI and metacognition in STEM education began to grow substantially after 2005, making it a meaningful cutoff for capturing contemporary developments.
The database search yielded 168 records, of which 33 duplicates were removed, leaving 135 unique studies for screening. Ninety-five were excluded at the title/abstract stage, and 40 full-text articles were assessed for eligibility. Sixteen full-texts were excluded—6 did not address AI and metacognition jointly, 4 were not STEM-related, 3 were non-empirical, and 3 for other reasons—resulting in 24 studies included in the systematic review.

2.3. Data Retrieval and Screening Reliability

A total of 119 documents were retrieved from Scopus and 49 from Web of Science. Following deduplication and file merging in RStudio (version 2025.05.1+513), a final dataset of 135 unique records was established for bibliometric analysis. These documents were saved in BibTeX and CSV formats for compatibility with VOSviewer (version 1.6.20) and the Biblioshiny interface in RStudio. Following retrieval and cleaning, the included studies were thematically coded. Thematic codes were developed through an iterative review process. Two researchers independently coded an initial subset of 10 studies to refine the codebook before applying it to the remaining articles. To ensure screening rigour, both reviewers independently evaluated all titles, abstracts, and full texts against the inclusion/exclusion criteria. Agreement was assessed using Cohen’s κ, showing substantial reliability (κ = 0.82 for title/abstract screening; κ = 0.87 for full-text). Discrepancies were resolved through discussion with a third reviewer. The final thematic categories are summarised in Table 1.
Table 1. Summary of Empirical Studies Integrating AI and Metacognition in STEM Education.

2.4. Bibliometric Analysis

Using VOSviewer and Biblioshiny, the study generated co-occurrence maps, keyword clustering, author collaboration networks, and thematic evolution diagrams. This helped to identify key themes, highly cited works, and emerging concepts in the literature related to AI, metacognition, and STEM education ().

2.5. Systematic Review and PRISMA Approach

Following PRISMA 2020 guidelines (), a subset of the 135 bibliometric records was systematically screened through the four phases—Identification, Screening, Eligibility, and Inclusion—to identify the 24 studies selected for qualitative synthesis (see Figure 1).
Figure 1. PRISMA 2020 flow diagram summarising the study-selection process. Note. This figure illustrates the systematic process used for identifying and screening, and includes studies that examined the facilitation of metacognition in AI-supported STEM classrooms.
These studies were analysed to explore how metacognitive strategies are integrated with AI tools in STEM education contexts. The review went beyond theoretical inquiry, capturing empirical insights on implementing such combinations in classroom practice. Key data points extracted from each study included: the theoretical or conceptual frameworks employed, the nature and purpose of metacognitive strategies used, the specific AI tools or systems applied (e.g., intelligent tutoring systems, chatbots, adaptive platforms), the STEM discipline and educational level (e.g., primary, secondary or tertiary), and the study’s alignment with human-centred or posthumanist paradigms (see Table 1 for the empirical studies integrating AI and metacognition in STEM education). In clarifying the posthumanist paradigm, we note that what is “shared” is not metacognition itself, but rather the regulation of learning experiences, where metacognitive processes are scaffolded by AI tools alongside human agency. The data were coded thematically using a structured Excel matrix, which enabled analysis of trends in theoretical orientation, the evolving relationship between AI and metacognition, and the pedagogical implications of these interactions across STEM learning environments.
Across the reviewed studies, AI applications supporting metacognition were most implemented in mathematics and science education, particularly in secondary and higher education contexts. Intelligent Tutoring Systems, learning analytics dashboards, and adaptive learning platforms were the most prevalent AI tools used to scaffold metacognitive regulation and awareness. Despite this progress, significant research gaps persist across STEM domains and various educational levels. Primary-school and early-childhood STEM education are scarcely represented, and few studies examine AI-driven metacognitive interventions in teacher-education or pre-service training. Similarly, engineering and technology subjects receive far less attention than mathematics and science. These imbalances suggest that current research is concentrated in only two of the four STEM areas, indicating the need for broader, domain-inclusive exploration of AI-mediated metacognition.

2.6. Summary of Findings

Analysis of the 24 reviewed studies reveals a clear dominance of human-centred approaches, with 23 studies explicitly framed within the context of human agency. In contrast, only one study () reflected a posthumanist orientation. Most interventions focused on planning, monitoring, and reflection strategies, particularly within mathematics and general STEM contexts. Secondary and tertiary education levels were the most represented, while primary-level studies remained underrepresented. These findings indicate that while AI tools are increasingly used to scaffold metacognition, theoretical perspectives remain largely traditional, highlighting an important gap for future exploration of hybrid or posthumanist models.

3. Results

The findings are organised and reported according to the structure of the research questions guiding this study.

3.1. RQ1: What Are the Publication Trends on Metacognition in STEM Education Between 2005 and 2025?

The publication output on metacognition in STEM education has steadily increased since 2005, with notable growth after 2015 (see Figure 2). The annual output grew significantly from a single in 2005, indicating a rising scholarly interest. The trend suggests a peak in recent years, which aligns with global attention on educational innovation, artificial intelligence, and learner autonomy.
Figure 2. Annual Publication Trends on Metacognition in STEM Education (2005–2025). Note. This figure displays the number of publications per year from 2005 to 2025 related to metacognition in STEM education. Data were retrieved from a bibliometric analysis of 135 documents using Bibliometrix and Biblioshiny.

3.2. RQ2: What Are the Most Frequently Occurring AI-Related Concepts and Tools in the Literature on Metacognition in STEM Education?

A keyword co-occurrence analysis was conducted to identify the most prominent AI-related concepts and tools within the literature on metacognition in STEM education. As shown in Table 2 (ranked by frequency) and visualised in Figure 3, the most frequently occurring AI-related term was learning analytics (n = 43), underscoring its centrality in supporting and monitoring self-regulated learning processes. Other high-ranking terms include learning systems (n = 25), artificial intelligence (n = 20), and e-learning (n = 17), suggesting widespread integration of digital and intelligent technologies in STEM education. Notably, emerging approaches such as adversarial machine learning and contrastive learning (each n = 11), along with generative AI (n = 10), reflect increasing interest in advanced AI models. Traditional applications, such as intelligent tutoring systems (n = 10), also remain relevant, particularly in fostering the use of metacognitive strategies. These findings illustrate a shift from established AI tools toward more sophisticated, data-driven educational technologies.
Table 2. Frequently Occurring AI-Related Concepts and Tools in Literature on Metacognition in STEM Education.
Figure 3. Keyword co-occurrence network map of AI-related terms. Note. This map was generated using VOSviewer based on a minimum occurrence threshold of 10. Node size indicates keyword frequency. Line thickness shows the strength of co-occurrence links between keywords.

3.3. RQ3: Which Journals, Authors, and Countries Contribute Most to the Literature on AI and Metacognition in STEM Education?

3.3.1. Journal Contributions to Research on AI and Metacognition in STEM Education

A source analysis was conducted to identify the key publication venues contributing to research on AI and metacognition in STEM education. Results show that the literature is published in peer-reviewed journals and major conference proceedings. As presented in Table 3, the Lecture Notes in Computer Science (LNCS) series led with 16 publications, reflecting strong contributions from computer science-driven education research. This is followed by British Journal of Educational Technology, Education and Information Technologies, and Frontiers in Education, each with five articles. Notable venues, such as Computers & Education, IEEE EDUCON, and the Frontiers in Education Conference (FIE), contributed four articles each, highlighting sustained academic and practitioner interest. Furthermore, journals such as Computer Applications in Engineering Education, Computers in Human Behaviour, and Educational Psychology Review each produced three articles, signalling a growing interdisciplinary engagement with AI-supported metacognitive research.
Table 3. Top 10 Publication Venues Contributing to Research on AI and Metacognition in STEM Education.

3.3.2. Author Contributions to Research on AI and Metacognition in STEM Education

A co-authorship and productivity analysis revealed leading contributors to the literature on AI and metacognition in STEM education. As shown in Table 4 and visualised in Figure 4, Roger Azevedo leads with eight publications and 399 citations, reflecting strong and sustained engagement. Dragan Gašević follows with six publications and the highest citation count (720), highlighting his significant influence in the field. Other frequently publishing authors include Guanhua Chen, Charles Xie, Wanli Xing, Juan Zheng, and Michelle Taub, each with six publications. The VOSviewer map in Figure 4 displays author co-authorship networks, where the largest connected cluster comprises six items, indicating limited but growing collaboration among core researchers. These findings suggest a concentrated yet expanding scholarly community contributing to AI-driven metacognitive research in STEM education.
Table 4. Top Contributing Authors to Research on AI and Metacognition in STEM Education.
Figure 4. Author Co-authorship Network on AI and Metacognition in STEM Education. Note. The network shows co-authorship patterns among the most productive authors in the field.

3.3.3. Country Contributions to Research on AI and Metacognition in STEM Education

A bibliometric analysis of 135 publications, as shown in Table 5, revealed that the United States is the leading contributor to research on AI and metacognition in STEM education, with 68 publications, followed by China (37), Germany (16), Australia (13), and Canada (11). Other active contributors include South Korea and Spain (10 each), while Finland, France, and Italy each produced five publications. These results, visualised in Figure 5, illustrate the global distribution of scientific productivity in this field.
Table 5. Top 10 Countries Contributing to Research on AI and Metacognition in STEM Education.
Figure 5. Country Scientific Production on AI and Metacognition in STEM Education. Note. Visualisation generated using Biblioshiny. The map illustrates the geographic distribution of publication volume by country, based on corresponding author affiliations.

3.4. RQ4: Which Theoretical Frameworks Are Most Frequently Associated with Studies on AI and Metacognition in STEM, and How Have These Evolved?

Across the 25 reviewed studies, as shown in Table 1, Self-Regulated Learning (SRL) Theory is the most frequently used framework, particularly between 2010 and 2020. Zimmerman’s model of SRL, focused on planning, monitoring, and reflection, underpins many studies involving AI tools like dashboards and adaptive systems (e.g., ; ). From 2017 onward, the literature shows greater theoretical diversity. Frameworks such as Socially Shared Metacognition (), Design-Based Research (), and Reflective Pedagogy () are responsive to more collaborative and technology-mediated environments. Recent models, including ALERT and AI Safety Design (; ), signal a shift toward system-level metacognition embedded in AI. This trend reflects a move from individual learner regulation toward co-regulation within human–AI systems, expanding theoretical focus to include design ethics and technological cognition (Table 1, Columns 8–9).

3.5. RQ5: How Have Keywords and Conceptual Language in the Literature Shifted from Human-Centred to Posthumanist Paradigms in the Context of AI and Metacognition?

A conceptual shift is equally observable in the keywords and ontologies used across the studies. Earlier research (e.g., ; ) emphasised human-centred constructs such as reflection, self-awareness, regulation, and teacher scaffolding. As shown in Table 1, Column 9, the discourse remained embedded within the learner’s cognitive boundaries and relied heavily on the teacher or instructional design as the primary agent orchestrating metacognitive growth. In contrast, more recent studies (e.g., ) reflect an emerging posthumanist orientation, where metacognition is enacted by humans and embedded within or distributed across technological systems. Here, metacognition is no longer exclusively a human trait but a designable and operational feature of AI systems. Terms such as system-level reflection, machine awareness, AI safety frameworks, and autonomous decision regulation suggest that agency is increasingly diffused across networks of human and non-human actors.
This evolution marks a significant paradigm shift: from instrumental AI serving human learning to entangled learning ecologies in which humans and machines co-participate in reflective and self-regulatory processes. This shift transforms the learner–AI relationship and challenges conventional boundaries of cognition, strategy, and educational agency in STEM contexts.

4. Discussion

This section interprets the key findings thematically, providing possible explanations that are supported by relevant theories and literature. It examines the evolution of research on AI and metacognition in STEM, highlighting trends, tools, contributors, theoretical shifts, and emerging paradigms.

4.1. Growing Research Attention on AI and Metacognition in STEM Education

The steady rise in research output on metacognition within STEM education, especially after 2015, may be explained by increasing global attention to learner autonomy, personalised education, and the digital transformation of classrooms. As AI technologies have become more accessible and widely adopted in education systems, scholars have naturally been drawn to investigate how these tools affect higher-order learning processes, such as metacognition. Furthermore, the widespread emphasis on 21st-century skills, such as critical thinking, problem-solving, and self-regulated learning, likely contributed to this surge in interest. Researchers have sought to understand not only how students learn, but also how they learn to learn, particularly within the complex and rapidly evolving contexts of science and mathematics. This pattern aligns well with ’s () Theory of Metacognition, which posits that metacognition is essential for effective learning. The increasing research interest reflects recognition of the importance of metacognitive knowledge and regulation (planning, monitoring, and evaluation) in enabling students to succeed in cognitively demanding STEM environments.
Furthermore, the systemic nature of this research growth resonates with GST, as it illustrates how evolving technological, pedagogical, and psychological subsystems collectively shape educational research trajectories (). The literature supports this trend. Studies have consistently shown that metacognitive strategies improve learning outcomes in mathematics and science (e.g., ; ; ). These findings justify the intensifying research interest and affirm that metacognition has become central in AI-permeated STEM education as a theoretical construct and instructional target.

4.2. Learning Analytics and AI Tools as Scaffolds for Metacognitive Regulation

The prominence of terms like learning analytics, learning systems, and intelligent tutoring systems suggests that much of the research on AI and metacognition in STEM education has focused on how digital tools can support learners in planning, monitoring, and evaluating their learning. This focus likely stems from the growing demand for scalable, data-driven approaches to support self-regulated learning. As classrooms become increasingly digitised and complex, educators and learners require systems that offer timely feedback, track cognitive engagement, and guide real-time decision-making processes. These tools are not just replacing traditional instruction; they act as cognitive partners that help externalise metacognitive processes.
This development aligns conceptually with the GST, emphasising the interdependence of components in complex systems (). AI tools are subsystems that interact with instructional design, learner behaviour, and cognitive demands to create adaptive learning environments. These environments promote metacognitive growth by dynamically responding to learner inputs, just as a well-functioning system adapts to maintain equilibrium. In this context, learning regulation and learning experiences, supported through metacognitive scaffolds, become a shared responsibility between human and machine components. This trend is well-supported in the literature. For example, () showed how analytics tools embedded in online learning platforms improved students’ metacognitive awareness by visualising their behaviours and providing reflective feedback. Similarly, () found that tools like Lumilo and smart glasses enhanced teachers’ and students’ abilities to monitor and regulate learning in real time. These studies confirm that AI-powered systems can scaffold key metacognitive functions, especially in STEM classrooms where abstract reasoning and problem-solving are central. The shift toward digital metacognitive support thus reflects both technological capability and pedagogical necessity.

4.3. Scholarly Ecosystem—Productive Journals, Authors, and Countries

The dominance of certain journals, authors, and countries in this field highlights the increasingly interdisciplinary and globally coordinated nature of research at the intersection of AI, metacognition, and STEM education. The leading role of venues like Lecture Notes in Computer Science (LNCS) and the British Journal of Educational Technology suggests that educational researchers and computer scientists are shaping the field. This may reflect a convergence of priorities: educators seek technological supports for metacognitive development, while technologists embed cognitive theories into intelligent systems. The strong presence of engineering and psychology journals further underscores the multidisciplinary character of the field.
Prominent scholars such as Roger Azevedo and Dragan Gašević, known for their work on self-regulated learning and learning analytics, contribute heavily to this discourse. Their sustained engagement reflects a maturing field in which theoretical development is closely tied to technological innovation. Countries like the United States and China dominate the landscape, likely due to their greater access to funding, infrastructure, and institutional support for integrating AI in education. The rise of authors from these countries also suggests that global leadership in educational AI research is deeply linked to national investments in AI development and STEM policy priorities. For example, in the United States, initiatives such as the National AI Research Institutes (established by the National Science Foundation in 2020) and the Federal STEM Education Strategic Plan (2018) have directed substantial funding toward integrating AI and STEM in education. Similarly, in China, the 2017 Next Generation Artificial Intelligence Development Plan explicitly prioritised AI in education, with national-level funding allocated to intelligent tutoring systems, adaptive learning platforms, and AI-supported STEM curricula. These sustained investments help explain the concentration of research leadership in these two countries and provide context for their dominant scholarly contributions.
This distributed scholarly activity is well-explained by GST, which frames educational research as a complex system shaped by interacting subsystems, including academic publishing ecosystems, funding environments, and disciplinary networks. Research clustering in particular regions and institutions reflects system-level feedback loops, where resource availability, research visibility, and citation impact reinforce continued output. These findings also align with trends in the literature. For instance, () and () note a sharp rise in AI-STEM scholarship over the last decade, particularly in technologically advanced regions. The field’s evolution is thus not only conceptual but also structurally driven by collaborations among productive scholars, high-impact journals, and innovation-focused countries. While this concentration suggests a strong knowledge base, it also points to the need for broader global engagement to ensure the inclusive development of AI tools and metacognitive frameworks across diverse STEM contexts.

4.4. Theoretical Evolution—From Individual Regulation to System-Level Metacognition

Theoretical frameworks used in the reviewed literature demonstrate a clear progression from traditional models of individual metacognitive control to more complex, distributed frameworks that reflect the evolving nature of learning environments. The early dominance of Self-Regulated Learning theory, particularly Zimmerman’s model, can be attributed to its strong alignment with foundational metacognitive processes, planning, monitoring, and evaluating, especially in the context of student autonomy (e.g., ; ). As AI technologies began to mediate more aspects of learning, researchers increasingly incorporated complementary frameworks, such as Socially Shared Metacognition (), Reflective Pedagogy (), and Design-Based Research (), to account for the co-regulatory and collaborative dimensions of learning. This evolution resonates with General Systems Theory, which views learning environments as dynamic networks of human and technological actors (, ). More recent work in complexity science () strengthens this interpretation by emphasising adaptive interactions within complex systems; the emergence of system-level metacognition can be understood as one such adaptive phenomenon. Only one reviewed study () explicitly adopted a posthumanist orientation, embedding metacognitive reflection at the system level within AI design. While this single case does not signal a widespread paradigm shift, it represents an important departure from human-centred assumptions. When viewed alongside the bibliometric evidence—such as emerging keywords like system-level reflection and machine awareness—this suggests the beginnings of a theoretical diversification that may open space for posthumanist framings in the future.

4.5. Conceptual Shifts from Human-Centred to Posthumanist Frames

The conceptual language across the corpus reveals both continuity and subtle change. Earlier studies emphasised human-centred constructs such as reflection, self-awareness, and teacher scaffolding (e.g., ; ; ), reinforcing the assumption that cognition resides solely within the learner. More recent contributions, though still few in number, introduce terms such as autonomous decision regulation and AI safety frameworks, which imply a broader distribution of agency across human and technological actors. Although the systematic review shows that explicit posthumanist framings are rare (1 of 24 studies; see ), these linguistic and conceptual signals in the broader bibliometric dataset are consistent with the posthumanist perspectives discussed by scholars such as () and Lim 2025. Together, they suggest a nascent posthumanist discourse. It would therefore be premature to describe a consolidated “trend,” but it is reasonable to note that posthumanist perspectives are emerging as an alternative orientation. This interpretation recognises the dominance of human-centred paradigms while also acknowledging the potential for hybrid human–AI framings to shape future theorisation and practice in STEM education.

4.6. Clarifying “Scaffolding” Versus “Sharing” Metacognition

Building on this shift toward posthumanist framing, it is necessary to clarify how AI’s role in supporting metacognition is conceptualised, particularly in distinguishing scaffolding from sharing. We recognise an important distinction between scaffolding and sharing metacognition. Scaffolding treats metacognition as a human capacity that can be supported and extended through prompts, feedback, or adaptive interventions, thereby allowing learners to retain agency and responsibility. Sharing, by contrast, distributes metacognitive functions between human and AI agents, where AI may compensate for weak regulation (e.g., automating monitoring or evaluation). While this reflects posthumanist views on hybrid intelligence, it risks limiting learners’ opportunities to develop independent strategies if applied uncritically. Both perspectives acknowledge AI’s active role, but their implications diverge: scaffolding supports gradual learner growth, whereas sharing risks substitution. Our review positions AI primarily as a scaffolding tool for metacognition to strengthen learners’ reflective capacities, while noting that sharing must be critically evaluated.

5. Conclusions

This study contributes to the understanding of how artificial intelligence (AI) tools intersect with metacognition in STEM education by combining bibliometric and systematic review approaches. The analysis demonstrates that publications in this field have increased steadily, with a growing shift toward integrating AI-supported systems such as intelligent tutoring, adaptive feedback, and learning analytics into instructional design. Most of the reviewed studies focused on mathematics and science education, reflecting the dominance of these disciplines in AI-assisted metacognitive research. Conceptually, the review identified that human-centred paradigms continue to guide most AI-metacognition studies. At the same time, only a few papers adopt posthumanist or hybrid perspectives that view AI as a cognitive partner rather than a tool. Despite these advances, methodological, theoretical, and ethical limitations remain. Few studies employed longitudinal or experimental research designs capable of identifying causal or sustained effects of AI-mediated metacognitive interventions. Research also remains uneven across STEM domains, with limited representation in technology and engineering education as well as in early-childhood and teacher-education contexts. Ethical considerations concerning data privacy, transparency, and learner agency are often underexplored, underscoring the need for guiding frameworks that balance innovation with responsibility.
In conclusion, this review proposes four specific recommendations for advancing the field. First, longitudinal and experimental studies should be prioritised to examine the long-term effects of AI-supported metacognitive interventions. Second, hybrid and posthumanist frameworks should be expanded to understand how human and machine cognition co-construct learning. Third, the development of explicit ethical frameworks is required to ensure transparency, fairness, and accountability in AI-metacognitive systems. Ultimately, future research should extend to underrepresented STEM domains—particularly technology, engineering, and early childhood education—to promote inclusivity and equity in AI-mediated learning research. Collectively, these directions outline a balanced agenda for strengthening the theoretical, methodological, and ethical foundations of AI-metacognition research in STEM education.

Supplementary Materials

The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/jintelligence13110148/s1. Table S1. Boolean Search Strings for Scopus and Web of Science Used in the Bibliometric–Systematic Review.

Funding

This research received no external funding.

Institutional Review Board Statement

Not applicable.

Data Availability Statement

No new data were created or analysed in this study.

Acknowledgments

The authors acknowledge the use of ChatGPT and Grammarly for proofreading and language enhancement during the preparation of this manuscript. The authors have reviewed and edited the output and take full responsibility for the content of this publication.

Conflicts of Interest

The authors declare that they have no conflicts of interest.

Abbreviations

The following abbreviations are used in this manuscript:
AIArtificial Intelligence
STEMScience, Technology, Engineering and Mathematics
PRISMAPreferred Reporting Items for Systematic Reviews and Meta-Analyses
ITSIntelligent Tutoring Systems
GSTGeneral Systems Theory
RQResearch Question(s)

References

  1. Abdelshiheed, Mark, John Wesley Hostetter, Tiffany Barnes, and Min Chi. 2023. Leveraging deep reinforcement learning for metacognitive interventions across intelligent tutoring systems. arXiv arXiv:2304.09821. [Google Scholar] [CrossRef]
  2. Adadan, Emine. 2020. Analysing the role of metacognitive awareness in pre-service chemistry teachers’ understanding of gas behavior in a multi-representational instruction setting. Journal of Research in Science Teaching 57: 253–78. [Google Scholar] [CrossRef]
  3. Ader, Engin. 2019. What would you demand beyond mathematics? Teachers’ promotion of students’ self-regulated learning and metacognition. ZDM Mathematics Education 51: 613–24. [Google Scholar] [CrossRef]
  4. Aghabeygi, Bahareh, and Sahar Khanjani. 2020. The effectiveness of metacognition therapy on the marital quality of women. Journal of Education and Health Promotion 9: 99. [Google Scholar] [CrossRef]
  5. Akturk, Ahmet Oguz, and Ismail Sahin. 2011. Literature review on metacognition and its measurement. ProcediaSocial and Behavioral Sciences 15: 3731–36. [Google Scholar] [CrossRef]
  6. Albus, Patrick, and Tina Seufert. 2025. Artificial intelligence in education: The importance of metacognitive monitoring. In The Impact of Artificial Intelligence on Societies. Studies in Neuroscience, Psychology and Behavioral Economics. Edited by Christian Montag and Raja Ali. Cham: Springer. [Google Scholar] [CrossRef]
  7. Aria, Massimo, and Corrado Cuccurullo. 2017. Bibliometrix: An R-tool for comprehensive science mapping analysis. Journal of Informetrics 11: 959–75. [Google Scholar] [CrossRef]
  8. Atchley, Paul, Hannah Pannell, Kaelyn Wofford, Michael Hopkins, and Ruth Ann Atchley. 2024. Human and AI collaboration in the higher education environment: Opportunities and concerns. Cognitive Research: Principles and Implications 9: 20. [Google Scholar] [CrossRef]
  9. Azevedo, Roger. 2020. Reflections on the field of metacognition: Issues, challenges, and opportunities. Metacognition and Learning 15: 91–98. [Google Scholar] [CrossRef]
  10. Azevedo, Roger, and Jennifer G. Cromley. 2004. Does training on self-regulated learning facilitate students’ learning with hypermedia? Journal of Educational Psychology 96: 523–35. [Google Scholar] [CrossRef]
  11. Baten, Elke, Magda Praet, and Annemie Desoete. 2017. The relevance and efficacy of metacognition for instructional design in the domain of mathematics. ZDM Mathematics Education 49: 613–23. [Google Scholar] [CrossRef]
  12. Ben-David, Adi, and Anat Zohar. 2009. Contributions of meta-strategic knowledge to scientific inquiry learning. International Journal of Science Education 31: 1657–82. [Google Scholar] [CrossRef]
  13. Berthold, Kirsten, Heidi Röder, Daniel Knörzer, Wolfgang Kessler, and Alexander Renkl. 2011. The double-edged effects of explanation prompts. Computers in Human Behavior 27: 69–75. [Google Scholar] [CrossRef]
  14. Bogdanović, Ivana Z., Dušica D. Rodić, Tamara N. Rončević, Jelena D. Stanisavljević, and Zekri A. M. Zouhor. 2022. The relationship between elementary students’ physics performance and metacognition regarding using modified Know-Want-Learn strategy. International Journal of Science and Mathematics Education 20: 1907–26. [Google Scholar] [CrossRef]
  15. Bond, Melissa, Hassan Khosravi, Maarten De Laat, Nina Bergdahl, Violeta Negrea, Emily Oxley, Phuong Pham, Sin Wang Chong, and George Siemens. 2024. A meta systematic review of artifcial intelligence in higher education: A call for increased ethics, collaboration, and rigour. International Journal of Educational Technology in Higher Education 21: 4. [Google Scholar] [CrossRef]
  16. Braad, Eelco, Nick Degens, Wolmet Barendregt, and Wijnand Ijsselsteijn. 2022. Improving metacognition through self-explication in a digital self-regulated learning tool. Educational Technology Research and Development 70: 2063–90. [Google Scholar] [CrossRef]
  17. Branigan, Heather E., and David I. Donaldson. 2019. Learning from learning logs: A case study of metacognition in the primary school classroom. British Educational Research Journal 45: 765–84. [Google Scholar] [CrossRef]
  18. Broadbent, Jaclyn, and Wendy L. Poon. 2015. Self-regulated learning strategies & academic achievement in online higher education learning environments: A systematic review. The Internet and Higher Education 27: 1–13. [Google Scholar] [CrossRef]
  19. Chardonnens, Sarah. 2025. Adapting educational practices for Generation Z: Integrating metacognitive strategies and artificial intelligence. Frontiers in Education 10: 1504726. [Google Scholar] [CrossRef]
  20. Chen, Ching-Huei, Tung-Kai Liu, and Kun Huang. 2021. Scaffolding vocational high school students’ computational thinking with cognitive and metacognitive prompts in learning about programmable logic controllers. Journal of Research on Technology in Education 55: 527–44. [Google Scholar] [CrossRef]
  21. Clark, Renee M., Rasim Guldiken, Autar Kaw, and Ozge Uyanik. 2025. The case for metacognition support in a flipped STEM course. International Journal of Mechanical Engineering Education 53: 720–48. [Google Scholar] [CrossRef]
  22. Clary, Renee M., Anastasia Elder, James Dunne, Svein Saebo, Debbie Beard, Charles Wax, and Deborah L. Tucker. 2018. Beyond the professional development academy: Teachers’ retention of discipline-specific science content knowledge throughout a 3-year mathematics and science partnership. School Science and Mathematics 118: 75–83. [Google Scholar] [CrossRef]
  23. Cuéllar-Rojas, Oscar-Andrés, Mauricio Hincapié-Montoya, Manuel Contero, and David Güemes-Castorena. 2022. Bibliometric analysis and systematic literature review of the intelligent tutoring systems. Frontiers in Education 7: 1047853. [Google Scholar] [CrossRef]
  24. Cukurova, Mutlu. 2024. The interplay of learning, analytics and artificial intelligence in education: A vision for hybrid intelligence. British Journal of Educational Technology 56: 469–88. [Google Scholar] [CrossRef]
  25. Desmarais, Michel C., and Ryan S. J. D. Baker. 2012. A review of recent advances in learner and skill modeling in intelligent learning environments. User Modeling and User-Adapted Interaction 22: 9–38. [Google Scholar] [CrossRef]
  26. Egara, Felix Oromena, and Mogege Mosimege. 2024a. Exploring the Integration of Artificial Intelligence-Based ChatGPT into Mathematics Instruction: Perceptions, Challenges, and Implications for Educators. Education Sciences 14: 742. [Google Scholar] [CrossRef]
  27. Egara, Felix Oromena, and Mogege Mosimege. 2024b. Effect of blended learning approach on secondary school learners’ mathematics achievement and retention. Education and Information Technologies 29: 19863–88. [Google Scholar] [CrossRef]
  28. Egara, Felix Oromena, and Mogege Mosimege. 2024c. Effect of Flipped Classroom Learning Approach on Mathematics Achievement and Interest among Secondary School Students. Education and Information Technologies 29: 8131–50. [Google Scholar] [CrossRef]
  29. Egara, Felix Oromena, Mogege Mosimege, and Moeketsi Mosia. 2025. Secondary school students’ perceptions of their usage of artificial intelligence-based ChatGPT in mathematics learning. Journal of Education 98: 124–46. [Google Scholar] [CrossRef]
  30. El Fathi, Tarik, Aouatif Saad, Hayat Larhzil, Driss Lamri, and El Mehdi Al Ibrahmi. 2025. Integrating generative AI into STEM education: Enhancing conceptual understanding, addressing misconceptions, and assessing student acceptance. Disciplinary and Interdisciplinary Science Education Research 7: 6. [Google Scholar] [CrossRef]
  31. Fatimah, Siti, Sarwi Sarwi, and Sri Haryani. 2024. Artificial intelligence in STEM education: A bibliometric analysis. International Journal of Multicultural and Multireligious Understanding 11: 186–200. [Google Scholar]
  32. Flavell, John H. 1979. Metacognition and cognitive monitoring: A new area of cognitive–developmental inquiry. American Psychologist 34: 906–11. [Google Scholar] [CrossRef]
  33. Fung, Chak Him, and Kin Keung Poon. 2021. Can dynamic activities boost mathematics understanding and metacognition? A case study on the limit of rational functions. International Journal of Mathematical Education in Science and Technology 52: 455–71. [Google Scholar] [CrossRef]
  34. Gamby, Sonja, and Christopher F. Bauer. 2022. Beyond “study skills”: A curriculum-embedded framework for metacognitive development in a college chemistry course. International Journal of STEM Education 9: 61. [Google Scholar] [CrossRef] [PubMed]
  35. González, Antonio, María-Victoria Carrera Fernández, and Paola-Verónica Paoloni. 2017. Hope and anxiety in physics class: Exploring their motivational antecedents and influence on metacognition and performance. Journal of Research in Science Teaching 54: 558–85. [Google Scholar] [CrossRef]
  36. Hidayat, Riyan, Hermandra Hermandra, Zetriuslita Zetriuslita, Shindy Lestari, and Hilman Qudratuddarsi. 2022. Achievement goals, metacognition and horizontal mathematisation: A mediational analysis. TEM Journal 11: 1537–46. [Google Scholar] [CrossRef]
  37. Holstein, Kenneth, and Vincent Aleven. 2021. Designing for human-AI complementarity in K-12 education. arXiv arXiv:2104.01266. [Google Scholar] [CrossRef]
  38. Jeong, Jeein, and Sangah Kim. 2025. Young children’s metacognition in problem-solving through question-asking. Journal of Experimental Child Psychology 219: 106207. [Google Scholar] [CrossRef] [PubMed]
  39. Jiang, Liming, Sheng Zhang, Xin Li, and Fang Luo. 2023. How grit influences high school students’ academic performance and the mediation effect of academic self-efficacy and cognitive learning strategies. Current Psychology 42: 94–103. [Google Scholar] [CrossRef]
  40. Kaplan, Avi. 2008. Clarifying metacognition, self-regulation, and self-regulated learning: What’s the purpose? Educational Psychology Review 20: 477–84. [Google Scholar] [CrossRef]
  41. Lavi, Rea, Gabriella Shwartz, and Yehudit Judy Dori. 2019. Metacognition in chemistry education: A literature review. Israel Journal of Chemistry 59: 583–97. [Google Scholar] [CrossRef]
  42. Levin, Ilya, Michal Marom, and Andrei Kojukhov. 2025. Rethinking AI in education: Highlighting the metacognitive challenge. Brain Broad Research in Artificial Intelligence and Neuroscience 16: 250. [Google Scholar] [CrossRef]
  43. Lim, Chaeyeon. 2025. DeBiasMe: De-biasing human-AI interactions with metacognitive AIED (AI in education) Interventions. In Proceedings of Workshop: AI Augmented Reasoning (CHI’25). New York: ACM, pp. 1–12. Available online: https://arxiv.org/pdf/2504.16770v1 (accessed on 11 July 2025).
  44. Lobczowski, Nikki G., Kayley Lyons, Jeffrey A. Greene, and Jacqueline E. McLaughlin. 2021. Socially shared metacognition in a project-based learning environment: A comparative case study. Learning, Culture and Social Interaction 30: 100543. [Google Scholar] [CrossRef]
  45. Lodhi, Shaouna. 2025. Integration of AI in STEM Education—Addressing Ethical Challenges in K-12 Settings. Master’s thesis, University of Minnesota, Minneapolis, MN, USA. Available online: https://conservancy.umn.edu/server/api/core/bitstreams/8cf44df5-26fe-423b-89f2-46f4bc68eb86/content (accessed on 17 July 2025).
  46. Margot, Kelly C., and Todd Kettler. 2019. Teachers’ perception of STEM integration and education: A systematic literature review. International Journal of STEM Education 6: 2. [Google Scholar] [CrossRef]
  47. Markandan, Nagalaxmy, Kamisah Osman, and Lilia Halim. 2022. Integrating Computational Thinking and Empowering Metacognitive Awareness in Stem Education. Frontiers in Psychology 9: 872593. [Google Scholar] [CrossRef]
  48. McCord, Rachel E., and Holly M. Matusovich. 2019. Naturalistic observations of metacognition in engineering: Using observational methods to study metacognitive engagement in engineering. Journal of Engineering Education 108: 218–42. [Google Scholar] [CrossRef]
  49. Medina, Melissa S., Ashley N. Castleberry, and Adam M. Persky. 2017. Strategies to improve learner metacognition in health professional education. American Journal of Pharmaceutical Education 81: 72–80. [Google Scholar] [CrossRef]
  50. Mitchell, Melanie. 2009. Complexity: A Guided Tour. Oxford: Oxford University Press. [Google Scholar]
  51. Mongeon, Philippe, and Adèle Paul-Hus. 2016. The journal coverage of Web of Science and Scopus: A comparative analysis. Scientometrics 106: 213–28. [Google Scholar] [CrossRef]
  52. Mosia, Moeketsi, and Felix Oromena Egara. 2025. Enhancing achievement and retention in circle geometry through digital storytelling for senior secondary learners. Education and Information Technologies, advance online publication. [Google Scholar] [CrossRef]
  53. Nelson, Thomas O., and Louis Narens. 1990. Metamemory: A theoretical framework and new findings. In The Psychology of Learning and Motivation. Edited by G. H. Bower. Cambridge: Academic Press, vol. 26, pp. 125–73. Available online: https://sites.socsci.uci.edu/~lnarens/1990/Nelson&Narens_Book_Chapter_1990.pdf (accessed on 15 June 2025).
  54. Oppong, Ernestina, Bruce M. Shore, and Krista R. Muis. 2019. Clarifying the connections among giftedness, metacognition, self-regulation, and self-regulated learning: Implications for theory and practice. Gifted Child Quarterly 63: 44–58. [Google Scholar] [CrossRef]
  55. Page, Matthew J., Joanne E. McKenzie, Patrick M. Bossuyt, Isabelle Boutron, Tammy C. Hoffmann, Cynthia D. Mulrow, Larissa Shamseer, Jennifer M. Tetzlaff, Elie A. Akl, Sue E. Brennan, and et al. 2021. The PRISMA 2020 statement: An updated guideline for reporting systematic reviews. BMJ 372: n71. [Google Scholar] [CrossRef]
  56. Pennequin, Valérie, Olivier Sorel, and Martial Mainguy. 2010. Metacognition, executive functions and aging: The effect of training in the use of metacognitive skills to solve mathematical word problems. Journal of Adult Development 17: 168–76. [Google Scholar] [CrossRef]
  57. Rajcoomar, Riona, Oarabile N. Morabe, and Bernard Breed. 2025. A socio-cultural model that promotes metacognition and creativity: An analysis of intercultural experiences within the metacognitive process in physical sciences classrooms. International Journal of Research in Education and Science 11: 319–40. [Google Scholar] [CrossRef]
  58. Roll, Ido, and Philip H. Winne. 2015. Understanding, evaluating, and supporting self-regulated learning using learning analytics. Journal of Learning Analytics 2: 7–12. [Google Scholar] [CrossRef]
  59. Roque Herrera, Yosbanys, Manuel Cañas Lucendo, Santiago Alonso García, and Alejandro Martínez Menéndez. 2025. Metacognition and academic performance in health sciences students. Interuniversity Journal of Teacher Training. Continuation of the Old Revista de Escuelas Normales 100: 87–103. [Google Scholar] [CrossRef]
  60. Schraw, Gregory, and David Moshman. 1995. Metacognitive Theories. Educational Psychology Review 7: 351–71. [Google Scholar] [CrossRef]
  61. Shields, Michelle, Grai Calabro, and Diana Selmeczy. 2024. Active help-seeking and metacognition interact in supporting children’s retention of science facts. Journal of Experimental Child Psychology 237: 105772. [Google Scholar] [CrossRef]
  62. Shneiderman, Ben. 2020. Human-Centered Artificial Intelligence: Three Fresh Ideas. AIS Transactions Human-Computer Interaction 12: 109–24. [Google Scholar] [CrossRef]
  63. Siegel, Marcelle A. 2012. Filling in the distance between us: Group metacognition during problem solving in a secondary education course. Journal of Science Education and Technology 21: 785–801. [Google Scholar] [CrossRef]
  64. Siegesmund, Amy. 2017. Using self-assessment to develop metacognition and self-regulated learners. FEMS Microbiology Letters 364: fnx096. [Google Scholar] [CrossRef]
  65. Smith, Julie M., and Rebecca Mancy. 2018. Exploring the relationship between metacognitive and collaborative talk during group mathematical problem-solving–what do we mean by collaborative metacognition? Research in Mathematics Education 20: 14–36. [Google Scholar] [CrossRef]
  66. Smith-Mutegi, Demetrice, Yoseph Mamo, Jinhee Kim, Helen Crompton, and Matthew McConnell. 2025. Perceptions of STEM education and artificial intelligence: A Twitter (X) sentiment analysis. International Journal of STEM Education 12: 9. [Google Scholar] [CrossRef]
  67. Stanton, Julie Dangremond, Amanda J. Sebesta, and John Dunlosky. 2021. Fostering Metacognition to Support Student Learning and Performance. CBE—Life Sciences Education 20: 1–7. [Google Scholar] [CrossRef]
  68. Swanson, Holly J., Adelola Ojutiku, and Bryan Dewsbury. 2024. The Impacts of an Academic Intervention Based in Metacognition on Academic Performance. Teaching & Learning Inquiry 12: 1–19. [Google Scholar] [CrossRef]
  69. Tak, Chan Choon, Hutkemri Zulnaidi, and Leong Kwan Eu. 2025. Mediating role of metacognitive awareness between attitude and mathematics reasoning in pre-service teachers. European Journal of Science and Mathematics Education 13: 90–102. [Google Scholar] [CrossRef] [PubMed]
  70. Tankelevitch, Lev, Viktor Kewenig, Auste Simkute, Ava Elizabeth Scott, Advait Sarkar, Abigail Sellen, and Sean Rintel. 2023. The metacognitive demands and opportunities of generative AI. arXiv arXiv:2312.10893. [Google Scholar] [CrossRef]
  71. Tanner, Kimberly D. 2017. Promoting student metacognition. CBE—Life Sciences Education 11: 113–20. [Google Scholar] [CrossRef] [PubMed]
  72. Taran, Elvina L., and Hansel Roy R. Nalla. 2019. Metacognitive awareness and attitudes toward problem-solving in science of senior high school students. Journal of Advances in Humanities and Social Sciences 5: 33–43. [Google Scholar] [CrossRef]
  73. Utami, Deka Dyah, Punaji Setyosari, Otto Fajarianto, Waras Kamdi, and Saida Ulfa. 2023. The correlation between metacognitive and problem solving skills among science students. EduLine: Journal of Education and Learning Innovation 3: 138–43. [Google Scholar] [CrossRef]
  74. Von Bertalanffy, Ludwig. 1950. An outline of general system theory. British Journal for the Philosophy of Science 1: 134–65. [Google Scholar] [CrossRef]
  75. Von Bertalanffy, Ludwig. 1968. General System Theory: Foundations, Development, Applications. Brooklyn: George Braziller. [Google Scholar]
  76. Walker, Peter B., Jonathan J. Haase, Melissa L. Mehalick, Christopher T. Steele, Dale W. Russell, and Ian N. Davidson. 2025. Harnessing metacognition for safe and responsible AI. Technologies 13: 107. [Google Scholar] [CrossRef]
  77. Wang, Karen D., Zhangyang Wu, L’Nard Tufts, Carl Wieman, Shima Salehi, and Nick Haber. 2024. Scaffold or crutch? Examining college students’ use and views of generative AI tools for STEM education. arXiv arXiv:2412.02653. [Google Scholar] [CrossRef]
  78. Willison, John, Claire Draper, Laura Fornarino, Menghua Li, Tala Sabri, Yan Shi, and Xinshuo Zhao. 2023. Metacognitively ALERT in science: Literature synthesis of a hierarchical framework for metacognition and preliminary evidence of its viability. Studies in Science Education 60: 153–89. [Google Scholar] [CrossRef]
  79. Yang, Kai-Lin. 2012. Structures of cognitive and metacognitive reading strategy use for reading comprehension of geometry proof. Educational Studies in Mathematics 80: 307–26. [Google Scholar] [CrossRef]
  80. Yin, Jiaqi, Yi Zhu, Tiong-Thye Goh, Wen Wu, and Yi Hu. 2024. Using Educational Chatbots with Metacognitive Feedback to Improve Science Learning. Applied Sciences 14: 9345. [Google Scholar] [CrossRef]
  81. Zepeda, Cristina D., Christina O. Hlutkowsky, Anne C. Partika, and Timothy J. Nokes-Malach. 2019. Identifying teachers’ supports of metacognition through classroom talk and its relation to growth in conceptual learning. Journal of Educational Psychology 111: 522–41. [Google Scholar] [CrossRef]
  82. Zhang, Hongmei, and Yanju Li. 2021. Integrating active learning activities and metacognition into STEM writing courses. Advances in Physiology Education 45: 742–48. [Google Scholar] [CrossRef]
  83. Zhang, Jing, Yu Zhou, Bin Jing, Zhongling Pi, and Hongliang Ma. 2024. Metacognition and mathematical modeling skills: The mediating roles of computational thinking in high school students. Journal of Intelligence 12: 55. [Google Scholar] [CrossRef]
  84. Zhao, Ningning, Xichun Teng, Wenting Li, Yanhua Li, Shuaiming Wang, Hongbo Wen, and Mengya Yi. 2019. A path model for metacognition and its relation to problem-solving strategies and achievement for different tasks. ZDM Mathematics Education 51: 641–53. [Google Scholar] [CrossRef]
  85. Zohar, Anat, and Sarit Barzilai. 2013. A review of research on metacognition in science education: Current and future directions. Studies in Science Education 49: 121–69. [Google Scholar] [CrossRef]
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Article Metrics

Citations

Article Access Statistics

Journal Statistics
Multiple requests from the same IP address are counted as one view.
Download PDF
An Analysis of Individual Differences in Within-Test Practice Effects in Progressive Matrices
Previous