Integrating immersive technologies, artificial emotional intelligence, and FabLabs in early-childhood STEAM education: a systematic review

Abstract

Introduction:

Early–childhood STEAM education increasingly incorporates immersive, affective, and maker–based technologies; however, existing evidence remains conceptually fragmented across physical, digital, and emotional domains. This study systematically synthesizes interdisciplinary research on Augmented and Virtual Reality (AR/VR), Artificial Emotional Intelligence (AEI), and FabLab–based interventions in early–childhood STEAM contexts.

Methods:

Following PRISMA guidelines, eligible empirical studies were analyzed through systematic review procedures, and a complementary meta-analytic synthesis was conducted to estimate the effects of these technologies on learning outcomes.

Results:

The findings indicate a statistically significant positive effect on learning outcomes. Immersive technologies primarily enhance conceptual understanding and motivation; AEI supports engagement and emotional regulation; and maker–based environments foster creativity, collaboration, and embodied problem–solving.

Discussion:

Despite these benefits, few interventions intentionally integrate these mediational domains. The study advances a Physical-Digital-Emotional (PDE) framework that reconceptualizes early–childhood STEAM learning as the coordinated alignment of embodied, cognitive, and affective processes. This integrative model strengthens the link between theory and practice and offers a developmentally grounded foundation for designing holistic STEAM environments in early childhood.

1 Introduction

STEAM education (Science, Technology, Engineering, Arts, and Mathematics) has gained increasing relevance in early-childhood settings as a pedagogical approach that promotes holistic, inquiry-based, and creative learning. At this developmental stage, STEAM emphasizes curiosity, exploration, and meaning-making through the integration of cognitive, socio-emotional, and psychomotor processes within playful and developmentally appropriate contexts (Garzón and Acevedo, 2019; Ng et al., 2022). Despite its growing adoption, early-childhood STEAM implementation often remains fragmented, relying on isolated disciplinary approaches or short-term technological interventions lacking theoretical integration and methodological coherence (Lampropoulos, 2024). Although positive effects on creativity, early scientific thinking, and engagement have been reported, heterogeneity in design and outcomes limits the consolidation of transferable evidence (Page et al., 2021).

Emerging technologies have increasingly been positioned as mediators of early STEAM learning. Immersive technologies, particularly Augmented Reality (AR) and Virtual Reality (VR), facilitate visualization of abstract phenomena and can enhance motivation and conceptual understanding when embedded in guided, age-appropriate designs (Billinghurst et al., 2001; Fokides and Samioti, 2024; Rapti et al., 2024; Pan and Melhem, 2024; Eryigit et al., 2025; Jiang et al., 2025). Artificial Emotional Intelligence (AEI), grounded in affective computing, enables adaptive feedback based on learners’ emotional states, supporting engagement and emotional regulation in early educational contexts (Picard, 1997; de Haas et al., 2021; Sánchez Fernández et al., 2025; Choi et al., 2025; Peleg and Levy, 2024). FabLabs and makerspaces contribute a tangible dimension through learning-by-making, fostering creativity, collaboration, and embodied problem-solving aligned with constructivist and constructionist traditions (Bagiati and Evangelou, 2015; Forbes et al., 2021; Barbosa et al., 2024; Videla-Reyes and Aros, 2025).

While these domains demonstrate complementary strengths, research typically examines them independently. This separation is particularly problematic in early childhood, where learning processes are inherently multidimensional and interdependent (Lampropoulos, 2024). A systematic synthesis capable of integrating physical, digital, and emotional mediations is therefore needed to clarify cumulative evidence and guide developmentally coherent practice. To address this gap, the present study systematically synthesizes empirical research on AR/VR, AEI, and FabLab-based interventions in early-childhood STEAM and incorporates a meta-analytic component to estimate the magnitude and consistency of reported effects. Building on the convergent patterns identified, the study advances an integrative Physical–Digital–Emotional (PDE) framework that reconceptualizes early-childhood STEAM as the coordinated alignment of embodied construction, immersive cognition, and adaptive affective scaffolding.

2 Conceptual background and analytical dimensions

Early-childhood STEAM education emphasizes holistic learning processes in which cognitive development, emotional engagement, and physical interaction are closely intertwined through play, exploration, and guided activity. Due to the interdisciplinary and developmental characteristics of this educational stage, isolated theoretical perspectives or technology-centered models are insufficient to explain how learning unfolds in technology-enhanced environments (Ng et al., 2022; Johnston et al., 2022; Leavy et al., 2023). Recent research therefore converges on integrative approaches that focus on how pedagogical design and technological mediation jointly shape early learning experiences (Wahyuningsih et al., 2020; Ng et al., 2022).

Across empirical studies and systematic reviews, three interrelated learning dimensions consistently emerge in early-childhood STEAM contexts: physical, digital, and emotional. These dimensions provide a concise analytical lens for synthesizing heterogeneous evidence and for examining how immersive technologies, Artificial Emotional Intelligence (AEI), and FabLab-based environments contribute to learning in developmentally appropriate ways (Lampropoulos, 2024; Jiang et al., 2025; Vistorte et al., 2024). Rather than functioning as independent domains, these dimensions interact dynamically during learning activities and are mediated by specific pedagogical configurations.

2.1 Physical dimension: learning through action and making

The physical dimension refers to embodied interaction, manipulation of materials, and learning through making. Grounded in constructivist and constructionist traditions, this dimension emphasizes the externalization of ideas through tangible artifacts and iterative design processes (Bagiati and Evangelou, 2015; Johnston et al., 2022; Leinonen et al., 2020). In early-childhood contexts, FabLabs and educational makerspaces operationalize this dimension by supporting hands-on exploration, collaboration, and creative problem-solving. Empirical studies associate such experiences with increased motivation, emerging agency, and positive attitudes toward science and engineering when activities are short, scaffolded, and aligned with curricular goals (Bagiati and Evangelou, 2015; Forbes et al., 2021; Peleg and Levy, 2024; Videla-Reyes and Aros, 2025).

2.2 Digital dimension: immersive visualization and experience

The digital dimension is primarily mediated through immersive technologies such as Augmented Reality (AR) and Virtual Reality (VR), which support visualization and experiential learning. AR enriches real-world contexts by overlaying digital information, preserving classroom interaction while enhancing conceptual understanding. VR enables immersive exploration of simulated environments, offering access to phenomena that are otherwise inaccessible (Billinghurst et al., 2001; Jiang et al., 2025; Sánchez Fernández et al., 2025). More broadly, metaverse-driven VR environments have also been explored as immersive systems for visualization, simulation, and process monitoring in complex digital contexts, reinforcing the relevance of VR as a mediational technology beyond conventional classroom settings (Alpala et al., 2026). Evidence indicates that both technologies can enhance engagement and learning outcomes when exposure is brief, structured, and pedagogically mediated, particularly in early-childhood settings where developmental constraints require careful design (Altmeyer et al., 2020; Criollo-C et al., 2024; Eryigit et al., 2025; Fokides and Samioti, 2024; Pan et al., 2021; Pan and Melhem, 2024; Rapti et al., 2024).

2.3 Emotional dimension: engagement and affective support

The emotional dimension addresses motivation, engagement, emotional regulation, and social interaction as central mediators of early learning. Artificial Emotional Intelligence (AEI), grounded in affective computing and educational AI, enables systems to respond adaptively to learners’ emotional cues by adjusting feedback, pacing, or task difficulty (Picard, 1997; Vistorte et al., 2024; Dignam et al., 2025). Studies involving social robots and affective agents suggest that AEI-supported interaction can enhance attention, engagement, and emotional self-regulation when integrated into guided activities under teacher supervision (de Haas et al., 2021; Sánchez Fernández et al., 2025; Choi et al., 2025; Lampropoulos, 2025). In early-childhood education, where emotional expression is often non-verbal and highly dynamic, this dimension plays a critical role in sustaining learning readiness and participation.

Together, these three dimensions capture the core mechanisms through which emerging technologies mediate early-childhood STEAM learning. This dimensional perspective provides the conceptual scaffold for the systematic review and meta-analysis presented in this study and underpins the integrative framework proposed in subsequent sections. Rather than offering an exhaustive theoretical exposition, this section functions as a focused analytical bridge linking evidence synthesis, quantitative analysis, and the proposed Physical–Digital–Emotional framework.

3 Methodology

3.1 Study design and review protocol

This study adopts a systematic review with quantitative meta-analysis, conducted in accordance with the PRISMA 2020 guidelines (Page et al., 2021). The review protocol was not prospectively registered. The choice of a systematic review responds to the fragmented and interdisciplinary nature of research on immersive technologies (AR/VR), Artificial Emotional Intelligence (AEI), and FabLab/Maker environments in early-childhood STEAM education. A meta-analysis was incorporated to statistically synthesize effect sizes from studies reporting comparable quantitative outcomes, thereby strengthening the level of evidence beyond narrative aggregation.

3.2 Information sources and search strategy

A systematic literature search was conducted between March and May 2025 across Scopus, Web of Science (Core Collection), IEEE Xplore, Dialnet, and Latindex. These databases were selected to ensure broad international coverage in education, technology, and interdisciplinary STEAM research, including both high-impact journals and relevant Ibero-American scholarship. The search strategy combined controlled vocabulary and natural language terms structured around four analytical dimensions: educational context, STEAM/STEM framework, early-childhood population, and focal technologies (AR/VR, Artificial Emotional Intelligence, and FabLab/maker environments).

A representative search query implemented in Scopus was as follows (adapted to the syntax of each database):

TITLE-ABS-KEY ((educat* OR school* OR kindergarten) AND (STEAM OR STEM) AND (child* OR early childhood OR primary) AND (augmented reality OR virtual reality OR AR OR VR OR immersive) AND (emotion* AI OR affective computing OR empathy) AND (FabLab OR makerspace OR digital fabrication OR maker education))

Iterative refinements incorporated conceptual variants (e.g., extended/mixed reality, social robots, affective tutors) and wildcard operators to maximize retrieval sensitivity. The temporal scope was limited to 2015–2025 to capture recent technological developments, while foundational pre-2015 reviews were selectively retained when theoretically relevant. No language restrictions were applied during the search stage; eligibility criteria were enforced during screening.

No formal backward snowballing procedure was conducted. However, the search strategy included database searching and complementary source screening, as reflected in the PRISMA flow diagram. All records identified through complementary sources were screened using the same eligibility criteria applied during the selection process.

3.3 Eligibility criteria and study selection (PRISMA)

The study selection process was conducted in strict accordance with the PRISMA 2020 guidelines to ensure transparency, replicability, and methodological rigor. Explicit inclusion and exclusion criteria were defined a priori and consistently applied throughout all screening stages.

3.3.1 Inclusion criteria

3.3.2 Exclusion criteria

• Studies were excluded if they met any of the following conditions:

• Targeted secondary, tertiary, or adult education contexts beyond the defined early-childhood and early primary scope;

• Were purely theoretical, conceptual, or speculative proposals without empirical implementation or evaluation (except for systematic reviews);

• Lacked sufficient methodological rigor, such as unclear research design, insufficient data reporting, or absence of outcome measures;

• Constituted duplicate publications identified across databases.

3.3.3 PRISMA selection outcomes

Application of the PRISMA 2020 protocol yielded a structured multi-stage screening process. A total of 395 records were initially identified across databases. After automated duplicate removal, 373 unique records remained for title and abstract screening, resulting in the exclusion of 145 studies. The remaining 228 records proceeded to full-text assessment, of which 46 articles were evaluated against predefined eligibility criteria using a standardized checklist.

During eligibility assessment, 22 studies were excluded due to language restrictions (Russian or Chinese), absence of the arts dimension (pure STEM focus), lack of empirical evaluation, or duplication of content.

A final set of 24 studies met all inclusion criteria. Of these, 17 empirical studies constituted the analytical corpus for qualitative synthesis and meta-analysis, while 7 systematic reviews and meta-analyses provided theoretical and contextual support for interpretation.

The complete selection process is illustrated in Figure 1, which presents the PRISMA 2020 flow diagram detailing identification, screening, eligibility, and inclusion stages.

Figure 1

Page et al., 2021).

3.4 Quality assessment of empirical studies

This three-layer approach reduces single-tool blind spots and enables a coherent appraisal across heterogeneous methods typical of early-childhood STEAM research.

3.4.1 Risk of bias assessment (RoB 2.0)

The RoB 2.0 tool (Sterne et al., 2019) was applied to empirical studies employing experimental or quasi-experimental designs, particularly those involving AR/VR interventions with comparison groups. Overall, the pattern indicates a low-to-moderate risk of bias, consistent with methodological constraints typical of classroom-based research.

The strongest domains were random sequence generation and selective outcome reporting, supporting the internal consistency of the quantitative evidence. Allocation concealment showed moderate limitations, while blinding emerged as the most vulnerable domain. This limitation is structurally inherent in educational interventions, where participants and teachers are aware of the technological condition (e.g., use of VR headsets, AR devices, or makerspace activities), potentially introducing performance or expectancy effects.

Table 1 summarizes the distribution of risk levels across assessed domains.

The overall low-to-moderate risk profile supports the inclusion of these studies in the meta-analysis. Nevertheless, the elevated risk observed in blinding should be acknowledged as a structural limitation inherent to educational interventions and interpreted cautiously, particularly in short-duration designs where expectancy or performance effects may be amplified. To account for residual heterogeneity and between-study variability, a random-effects model was employed in the quantitative synthesis.

3.4.2 Methodological coherence (MMAT)

The Mixed Methods Appraisal Tool (MMAT; Hong et al., 2018) was applied to qualitative, case-study, and mixed-methods studies within the empirical corpus, particularly those examining AEI/robotics and FabLab/Maker interventions, where contextual implementation is central.

Overall, the corpus demonstrates strong methodological coherence. All studies clearly articulated research objectives, and a high proportion exhibited alignment between reported findings and stated conclusions, supporting the credibility of the qualitative synthesis.

Table 2 presents the distribution of methodological quality across assessed MMAT dimensions.

The strongest dimensions—objective clarity and results–conclusion coherence—support the robustness of contextual interpretations presented in the narrative synthesis. Moderate limitations were observed in data triangulation and analytical transparency, indicating that some studies provide limited detail regarding integration of multiple data sources or explicit analytical procedures.

These constraints are particularly relevant in FabLab/Maker research, where learning outcomes are highly dependent on facilitation style and task configuration. Accordingly, findings derived from single-context implementations should be interpreted cautiously and not generalized beyond comparable pedagogical settings.

3.4.3 Overall strength of evidence (education-adapted GRADE)

To provide an article-level confidence statement aligned with high-impact systematic review standards, an education-adapted GRADE approach (Guyatt et al., 2011) was applied. Confidence was synthesized across five domains: risk of bias (RoB 2.0 + MMAT triangulation), consistency of findings, directness to early-childhood STEAM contexts, precision of estimates, and potential publication bias. The overall evidence assessment based on the education-adapted GRADE framework is summarized in Table 3.

Across domains, the evidence base demonstrates moderate-to-high overall confidence. AR/VR interventions show the most stable quantitative patterns, particularly for motivational and cognitive outcomes when pedagogically scaffolded. AEI and robotics interventions present consistent socio-emotional and engagement-related benefits, although effect magnitude appears sensitive to interaction design and adult mediation.

FabLab/Maker evidence remains credible but comparatively context-dependent, with fewer replicated experimental designs and greater reliance on case-based implementations. Consequently, confidence in these findings is moderate and contingent upon implementation quality, facilitation style, and curricular integration.

Taken together, the adapted GRADE synthesis supports the inclusion of these domains in an integrative framework while reinforcing the importance of developmental alignment and pedagogical mediation as primary determinants of effectiveness rather than technology type alone.

3.5 Data extraction and analytical corpus

Table 4 served as the operational core of the data extraction process. Its tabular design enabled the grouping of studies by technology and type of analyzed variable, providing a cross-sectional overview of the evidence. This procedure facilitated the development of the narrative synthesis and the selection of the 10 quantitative studies suitable for the meta-analysis.

3.6 Meta-analysis procedure

From the 17 empirical studies comprising the analytical corpus (Table 4), 10 met the criteria for quantitative comparability and were included in the meta-analysis. These studies contributed 10 effect-size estimates distributed across the three technological domains analyzed in this review: AR/VR, AEI/affective robotics, and FabLab/Maker environments. Studies were eligible if they: (i) reported quantitative outcomes related to learning, motivation, creativity, or socio-emotional variables; (ii) employed experimental, quasi-experimental, comparative, or pre–post designs; and (iii) provided sufficient statistical information to compute standardized effect sizes, including means, standard deviations, confidence intervals, or equivalent statistics.

The meta-analysis was therefore based on analytically comparable effect-size estimates rather than on a simple aggregation of participant counts, given the methodological heterogeneity of the included studies and the diversity of reported outcome measures. This approach allowed the quantitative synthesis to preserve comparability across technological domains while remaining aligned with the evidence extracted from the studies summarized in Table 4.

3.6.1 Effect size estimation

Effect sizes were calculated using Hedges’ g, which adjusts for small-sample bias and is therefore appropriate for early-childhood educational research contexts. When not explicitly reported, effect sizes were derived from published descriptive statistics following standard meta-analytic procedures. Magnitudes were interpreted using conventional benchmarks (0.20 = small; 0.50 = medium; ≥0.80 = large).

3.6.2 Meta-analytic model and heterogeneity

Given expected variation across technology type, intervention duration, educational context, and outcome measures, a random-effects model (DerSimonian–Laird estimator) was applied. This approach assumes that true effects vary across studies and provides more conservative and generalizable estimates than fixed-effect models. Heterogeneity was assessed using Cochran's Q and the I² index, interpreted as ≤30% (low), ∼50% (moderate), and ≥75% (high). Moderate-to-high heterogeneity was observed (I² ≈ 63%), reflecting substantive diversity in technological mediation and pedagogical implementation rather than methodological inconsistency.

3.7 Subgroup analyses

4 Results

4.1 Descriptive analysis of studies by technology

The final analytical corpus comprised 17 empirical studies selected through the PRISMA-based procedure described in the methodology section. These studies represent diverse geographic contexts across Europe, Asia, the Middle East, and Latin America, providing a cross-regional perspective on the implementation of emerging technologies in early-childhood STEAM education. While this diversity enhances contextual breadth and external relevance, it also introduces variability, which is addressed through subgroup analyses and random-effects meta-analytic modeling.

Most studies were conducted in preschool settings, with some extending into early primary education (Grades 1–3), reflecting increasing integration of technology-mediated STEAM at earlier developmental stages. Across the corpus, outcome variables cluster into four interconnected domains: (1) cognitive learning outcomes (e.g., conceptual understanding, literacy, spatial reasoning); (2) motivational and attitudinal variables (e.g., engagement, enjoyment, sustained attention); (3) socio-emotional regulation and interaction quality; and (4) collaborative and creative competencies, particularly in maker-based environments.

Across technological categories, interventions combining active participation, immersive or tangible engagement, and pedagogical scaffolding tend to report positive outcomes across multiple domains. This descriptive overview provides the empirical basis for the subsequent technology-specific synthesis and the quantitative meta-analysis, both derived exclusively from the studies summarized in Table 4.

4.2 Descriptive results by technological domain

4.2.1 AR/VR (immersive technologies)

Immersive technologies—Augmented Reality (AR) and Virtual Reality (VR)—constitute the largest subgroup within the empirical corpus, accounting for 7 of the 17 studies (41%) included in the descriptive and quantitative synthesis. The AR/VR empirical evidence included in the review is summarized in Table 5. These studies were conducted across diverse geographic and educational contexts and collectively provide robust evidence on the pedagogical potential of immersive environments in early-childhood STEAM education. Overall, the AR/VR interventions consistently report positive effects on both cognitive learning outcomes and motivational variables, with effect directions uniformly favoring technology-enhanced conditions over traditional instructional approaches.

4.2.1.1 Analytical synthesis

The evidence indicates complementary roles for AR and VR. AR enhances learning by overlaying interactive digital elements onto real-world materials, supporting spatial reasoning, literacy development, geometry, and environmental understanding while preserving classroom interaction. VR, in contrast, emphasizes immersive exploration, facilitating conceptual understanding and engagement in abstract or otherwise inaccessible contexts.

A recurrent finding across studies is that effectiveness depends on structured, multi-session, curriculum-aligned implementation rather than isolated or novelty-driven use. When pedagogically embedded and developmentally appropriate, immersive technologies demonstrate consistent benefits across cognitive and motivational domains, reinforcing their contribution to the overall positive effects observed in the quantitative synthesis.

4.2.2 Artificial emotional intelligence (AEI)

Artificial Emotional Intelligence (AEI) constitutes a distinct and emerging domain within early-childhood STEAM education, represented by 5 of the 17 empirical studies (29%) included in the review. The empirical evidence on artificial emotional intelligence interventions is summarized in Table 6. These studies focus on emotion-aware educational agents, such as social robots, affective tutors, and adaptive robotic systems, designed to perceive and respond to children's emotional and behavioral cues during learning activities. Overall, the evidence indicates moderate but highly consistent positive effects of AEI-based interventions, particularly in socio-emotional and engagement-related outcomes, which are critical at early developmental stages.

4.2.2.1 Analytical synthesis

Across the five AEI-focused studies, consistent improvements are observed in attention, emotional regulation, and sustained engagement, highlighting AEI's role at the intersection of cognition and affect in early-childhood STEAM learning. Unlike AR/VR interventions, which primarily address cognitive and perceptual outcomes, AEI-based approaches target emotional processes that influence learning readiness and persistence, with particularly strong benefits for children exhibiting lower sociability or higher frustration sensitivity. Social robots and affective tutors that adapt feedback based on interaction cues contribute to reduced disengagement, improved emotional self-regulation, and sustained task focus, indicating their potential to support inclusive educational practices. AEI systems also demonstrate high levels of acceptance among children, who frequently attribute social presence or empathetic qualities to these agents, thereby enhancing motivation when interactions are appropriately scaffolded and supervised by teachers. Overall, AEI functions as a complementary mediating layer that strengthens personalization and emotional support within STEAM environments, consistent with the moderate-to-large pooled effects identified in the quantitative meta-analysis.

4.2.3 Fablabs and maker environments

FabLabs and Maker-based learning environments represent the third major technological domain analyzed in this review, accounting for 5 of the 17 empirical studies (29%). The empirical evidence on FabLab and Maker-based environments is summarized in Table 7. These studies focus on hands-on, fabrication-oriented STEAM learning, in which children engage in the design, construction, testing, and refinement of tangible artifacts using age-appropriate tools and materials. Overall, the empirical evidence highlights the distinctive contribution of FabLab/Maker environments to the development of creative, collaborative, and process-oriented competencies, which are central to the STEAM educational philosophy.

4.2.3.1 Analytical synthesis

Across the FabLab/Maker studies, a consistent emphasis is placed on learning through making, with outcomes extending beyond conventional academic metrics. Children participating in maker-oriented activities demonstrate increased autonomy, initiative, and emerging leadership, as well as enhanced iterative problem-solving skills through cycles of trial, error, and refinement. A defining characteristic of FabLab/Maker environments is their reliance on collaborative learning dynamics. Studies consistently report improvements in peer interaction, shared decision-making, and collective troubleshooting, particularly when projects are structured around common goals and tangible outputs. These environments also promote embodied and kinesthetic learning, which is especially beneficial for young learners with diverse learning styles.

However, compared with AR/VR and AEI interventions, FabLab/Maker studies exhibit greater variability in outcomes. This variability is largely attributable to the central mediating role of teachers. The empirical evidence indicates that pedagogical vision, scaffolding strategies, and classroom management practices strongly shape the depth and quality of learning that occurs in makerspaces. When teacher guidance is intentional and aligned with curricular objectives, maker activities lead to richer learning experiences; conversely, limited scaffolding or underutilization of available technologies can constrain learning potential. Additionally, several studies highlight equity-related considerations, including gender differences in tool appropriation and leadership roles. These findings underscore the importance of intentional design choices, such as rotating roles and non-stereotyped task assignments, to ensure inclusive participation.

4.3 Meta-analysis results

To complement the descriptive synthesis presented earlier, a quantitative meta-analysis was conducted to estimate overall and technology-specific effects of emerging technologies on early-childhood STEAM learning outcomes. This section integrates statistical results with analytical interpretation in a coherent format suitable for scholarly reporting.

4.3.1 Overall meta-analytic effect

From the 17 empirical studies identified through the PRISMA selection process and summarized in Table 4, 10 met the criteria for quantitative comparability and were included in the meta-analysis. These studies employed experimental, quasi-experimental, comparative, or pre–post designs and provided sufficient statistical information to compute standardized effect sizes. The quantitative synthesis was based on 10 analytically comparable effect-size estimates rather than on a direct aggregation of participant counts. The meta-analysis summary of effects by technological subgroup is presented in Table 8.

Figure 2 presents individual study effect sizes (Hedges’ g) with 95% confidence intervals, along with the pooled estimate calculated using the DerSimonian–Laird random-effects model. The diamond represents the overall combined effect (g = 0.66), and the vertical reference line denotes the null effect (g = 0).

Figure 2

4.3.2 Subgroup analysis by technological domain

To further interpret sources of variability and to ensure alignment between quantitative findings and the descriptive synthesis, subgroup analyses were conducted by technological domain: AR/VR, Artificial Emotional Intelligence (AEI), and FabLab/Maker environments. Each study included in the meta-analysis contributed one analytically comparable effect-size estimate; therefore, the number of studies corresponds to the number of effect-size estimates included in each subgroup.

The subgroup analyses indicate variation in effect magnitude across technological domains. AR/VR interventions demonstrate the largest pooled effect (g = 0.78), suggesting comparatively strong associations with learning-related outcomes. The moderate heterogeneity observed in this subgroup likely reflects variation in implementation format, duration, and outcome focus rather than inconsistency in direction of effects.

AEI-based interventions yield a moderate-to-large pooled effect (g = 0.65) with comparatively lower heterogeneity (I² = 47%). This pattern suggests more stable effects across contexts, particularly for outcomes related to engagement, attention, and emotional regulation. The narrower variability may reflect the focused functional role of emotion-aware systems within guided classroom activities.

FabLab/Maker environments show a moderate pooled effect (g = 0.54), accompanied by the highest heterogeneity (I² = 63%). This dispersion is consistent with the contextual sensitivity of maker-based pedagogies, where instructional scaffolding, facilitation style, and task design substantially mediate outcomes.

Figure 3 presents pooled effect sizes and confidence intervals across technological domains. The visual comparison highlights both the comparatively larger magnitude associated with immersive technologies and the greater dispersion observed in maker-based interventions.

Figure 3

4.4 Subgroup analyses

To further interpret the robustness and generalizability of the meta-analytic findings, additional subgroup analyses were conducted based on age range, geographic region, and type of outcome evaluated. These analyses provide a deeper understanding of how contextual and developmental factors influence the effectiveness of technology-enhanced STEAM interventions in early-childhood education.

4.4.1 Subgroup analysis by age range

The analysis by educational stage revealed comparable effect sizes between preschool education (≤5 years) and early primary education (6–8 years). Pooled effects were approximately g ≈ 0.60 for preschool-aged children and g ≈ 0.65 for early primary students. This similarity indicates that the effectiveness of immersive technologies, AEI, and FabLab/Maker environments is not confined to a single early developmental stage, provided that interventions are appropriately adapted to children's cognitive, emotional, and motor capabilities.

These findings suggest that early exposure to STEAM-oriented technologies can be pedagogically beneficial across the transition from preschool to primary education, supporting continuity in learning trajectories rather than isolated, age-specific effects.

4.4.2 Subgroup analysis by geographic region

When studies were grouped by geographic context, a non-significant trend toward larger effect sizes was observed in Asian educational contexts (g ≈ 0.70) compared with Western contexts (g ≈ 0.55). Although this difference did not reach statistical significance, it may reflect contextual factors, such as variations in baseline instructional practices, curricular integration of technology, or teacher familiarity with structured technology-based interventions.

Importantly, the absence of statistically significant regional differences suggests that the positive effects of technology-enhanced STEAM education are broadly transferable across cultural and educational systems, reinforcing the global relevance of the findings.

4.4.3 Subgroup analysis by outcome type

The comparison between cognitive–academic outcomes (e.g., learning performance, comprehension, conceptual understanding) and creative–attitudinal outcomes (e.g., motivation, creativity, engagement, attitudes toward learning) revealed similar pooled effect sizes, both close to g ≈ 0.60. This result indicates that the analyzed interventions do not privilege one outcome domain at the expense of another.

Rather, technology-enhanced STEAM approaches appear to support multidimensional learning gains, simultaneously strengthening academic learning and key dispositions such as curiosity, persistence, and creative engagement—core objectives of the STEAM educational paradigm.

5 Proposed hybrid framework (physical–digital–emotional)

The meta-analysis findings indicate a statistically significant overall effect of immersive, affective, and maker-based interventions in early-childhood STEAM contexts, with differentiated contributions across technological domains. AR/VR interventions show strong cognitive–motivational effects; AI-based systems show consistent influence on engagement and emotional regulation; and FabLab environments promote embodied learning, collaboration, and creative agency. However, the quantitative synthesis also reveals that these approaches are often implemented in isolation, which limits multidimensional integration.

In response to this empirical pattern, we propose a Physical–Digital–Emotional (PDE) hybrid framework that reconceptualizes early-childhood STEAM learning as the coordinated activation of embodied construction, immersive cognition, and affective mediation. Rather than introducing new technologies, the framework reorganizes the convergent evidence identified in the meta-analysis into a unified structural model grounded in developmental coherence (Lampropoulos, 2024; Leavy et al., 2023).

5.1 Justification for a physical–digital–emotional framework (hands–head–heart)

The synthesis of 17 empirical studies reveals that early-childhood STEAM interventions tend to privilege one dominant mediating domain. Immersive technologies primarily enhance visualization and motivation; AEI interventions focus on emotional scaffolding; and maker environments emphasize tangible construction and creative agency. While each domain yields positive effects, the absence of intentional integration limits the developmental scope of interventions.

5.2 Conceptual foundations and learning dimensions (physical–digital–emotional)

The dimensions are analytically distinguishable but empirically interrelated, as suggested by the overlapping learning gains observed across studies.

5.3 Toward an integrative perspective: From fragmentation to convergence

The quantitative synthesis demonstrates that while individual technological domains produce moderate-to-large effects, few studies deliberately integrate physical, digital, and emotional mediation within a single pedagogical design. This structural separation suggests that current practice underutilizes potential cross-dimensional reinforcement.

Figure 4 illustrates the conceptual convergence of the three mediational domains. The model emphasizes that holistic early-childhood STEAM learning emerges from the intersection of embodied, cognitive, and affective processes rather than from technological immersion alone.

Figure 4

5.4 Framework structure: triangle model and pedagogical foundation

Figure 5

The triangle rests upon a foundational pedagogical base encompassing curriculum alignment, teacher mediation, and socio-cultural context. Technological affordances acquire educational meaning only within guided instructional ecologies; technology functions as mediated infrastructure rather than autonomous driver of learning.

5.4.1 Operational implications of the framework

Structurally, the framework implies a balanced configuration of learning environments integrating physical, digital, and emotional mediation.

6 Discussion

This systematic review and meta-analysis provides a consolidated perspective on how immersive technologies (AR/VR), Artificial Emotional Intelligence (AEI), and FabLab/Maker environments are applied in early-childhood STEAM education. Consistent with prior systematic reviews, the findings indicate that emerging technologies can positively influence early learning when they are developmentally appropriate and pedagogically mediated (Lampropoulos, 2024; Leavy et al., 2023; Wahyuningsih et al., 2020). Beyond confirming these trends, this study extends previous work by quantitatively synthesizing effects across multiple technological domains and by examining complementarities among physical, digital, and emotional dimensions within a unified analytical perspective.

The descriptive and meta-analytic findings show differentiated but complementary contributions across technological domains. AR/VR interventions appear particularly effective in supporting motivation, engagement, and conceptual understanding, especially when addressing abstract or spatial phenomena (Jiang et al., 2025; Pan and Melhem, 2024; Rapti et al., 2024; Khan Soomro et al., 2025; Zhang and Wang, 2021). AEI-based interventions contribute mainly to attention, emotional regulation, and sustained engagement, supporting evidence from affective robotics and AI-mediated learning environments (Choi et al., 2025; de Haas et al., 2021; Masturoh et al., 2024; Sánchez Fernández et al., 2025; Vistorte et al., 2024). FabLab/Maker environments show moderate but relevant effects on collaboration, active learning, creativity, and problem solving (Bagiati and Evangelou, 2015; Forbes et al., 2021; Leinonen et al., 2020; Peleg and Levy, 2024; Videla-Reyes and Aros, 2025). These results suggest that the three technological families should not be interpreted as competing instructional approaches, but as complementary mediational resources for early-childhood STEAM learning.

A central contribution of this study is the identification of persistent fragmentation in the literature. Most empirical interventions address immersive visualization, emotional support, or physical making in isolation, despite the multidimensional nature of early-childhood learning, where cognitive, affective, and psychomotor processes are closely interdependent (Chen, 2025; Garzón and Acevedo, 2019; Lampropoulos, 2024). The proposed Physical–Digital–Emotional (PDE) framework responds to this gap by organizing the evidence into a child-centered model that connects embodied construction, immersive cognition, and affective scaffolding within coherent pedagogical designs.

Methodologically, the study reinforces the need for rigorous and transparent evidence synthesis in early-childhood STEAM research. By combining systematic qualitative synthesis with meta-analysis, the findings suggest that educational effectiveness depends less on the specific technology employed and more on instructional design, adult mediation, developmental alignment, and implementation context (de Haas et al., 2021; Johnston et al., 2022; Leavy et al., 2023; Page et al., 2021). Therefore, the reported effects should be interpreted not as isolated technological advantages, but as outcomes shaped by the interaction between pedagogical design, learner characteristics, and contextual conditions.

A further practical implication concerns the unequal conditions under which immersive technologies, AEI systems, and FabLab/Maker environments can be implemented. In developing countries and resource-limited schools, adoption may be constrained by equipment costs, connectivity gaps, maintenance requirements, technical support needs, and limited institutional capacity. These constraints are especially relevant for VR devices, robotics platforms, affective computing systems, and digital fabrication tools, which require not only initial investment but also continuous updating, safe-use protocols, and sustained curricular integration.

Teacher readiness also emerges as a decisive condition for effective implementation. The findings indicate that technology type alone does not guarantee educational value; rather, outcomes depend on intentional scaffolding, classroom management, curriculum alignment, and teachers’ capacity to mediate children's physical, digital, and emotional engagement. In low-resource contexts, scalable implementation should therefore prioritize low-cost AR resources, shared or mobile maker kits, progressive teacher professional development, gradual adoption models, and hybrid pedagogical designs that do not rely exclusively on high-end infrastructure. Accordingly, the PDE framework should be understood not as a technology-intensive prescription, but as a flexible design logic adaptable to different institutional, cultural, and infrastructural conditions.

Overall, these findings suggest that early-childhood STEAM education should move beyond isolated technological adoption toward integrated pedagogical ecosystems. The value of the PDE framework lies in coordinating the physical, digital, and emotional dimensions of learning within developmentally appropriate experiences. Future research should examine how these mediations can be combined, sustained, and scaled across diverse educational contexts.

7 Conclusions

This systematic review and meta-analysis synthesized evidence on the integration of immersive technologies, Artificial Emotional Intelligence, and FabLab/Maker environments in early-childhood STEAM education. The findings confirm that technology-enhanced interventions can support meaningful learning outcomes when they are developmentally aligned, pedagogically mediated, and embedded within coherent instructional designs.

The main contribution of this study lies in moving beyond a technology-by-technology interpretation of the evidence. Although AR/VR, AEI, and FabLab/Maker environments show differentiated educational contributions, the reviewed literature reveals that these domains are still rarely integrated within unified pedagogical models. To address this gap, the proposed Physical–Digital–Emotional (PDE) framework offers an evidence-informed structure for connecting embodied making, immersive cognition, and affective scaffolding in early-childhood STEAM learning.

The study also reinforces the need to interpret emerging technologies not as autonomous drivers of educational improvement, but as mediational tools whose value depends on teacher guidance, curricular alignment, ethical use, and contextual feasibility. This point is especially relevant for resource-constrained educational settings, where implementation must consider infrastructure, cost, teacher readiness, and scalability.

Future research should prioritize longitudinal and large-scale studies capable of evaluating sustained learning effects, integrative pedagogical designs, and the practical adaptability of the PDE framework across diverse educational contexts. In this sense, the framework provides a foundation for designing more balanced, inclusive, and emotionally responsive STEAM environments for young learners.

References

• 1

  AlpalaL. O.Quiroga-ParraD. J.MontalvoF. J.PozoE. J.TorresJ. C. (2026). Towards industry 4.0: a metaverse-driven approach to smart factory process monitoring using virtual reality. Eng. Sci.40, 2097. 10.30919/es2097

  • CrossRef
  • Google Scholar
  • View reference in article

• 2

  AltmeyerK.KappS.TheesM.MaloneS.KuhnJ.BrünkenR. (2020). The use of augmented reality to foster conceptual knowledge acquisition in STEM laboratory courses—theoretical background and empirical results. Br. J. Educ. Technol.51 (3), 611–628. 10.1111/bjet.12900

  • CrossRef
  • Google Scholar
  • View reference in article

• 3

  BagiatiA.EvangelouD. (2015). Engineering curriculum in the preschool classroom: the teacher's experience. Eur. Early Child. Educ. Res. J.23 (1), 112–128. 10.1080/1350293X.2014.991099

  • CrossRef
  • Google Scholar
  • View reference in article

• 4

  BarbosaA.ValeI.AlvarengaD. (2024). The use of tinkercad and 3D printing in interdisciplinary STEAM education: a focus on engineering design. STEM Educ.4 (3), 222–246. 10.3934/steme.2024014

  • CrossRef
  • Google Scholar
  • View reference in article

• 5

  BillinghurstM.KatoH.PoupyrevI. (2001). The MagicBook: moving seamlessly between reality and virtuality. IEEE. Comput. Graph. Appl.21 (3), 6–8. 10.1109/38.920621

  • CrossRef
  • Google Scholar
  • View reference in article

• 6

  ChenJ. (2025). Exploring the role of artificial intelligence in personalized learning experiences in early childhood education to improve children’s social skills. Lect. Notes Educ. Psychol. Public Med.87, 121–132. 10.54254/2753-7048/2025.bo24154

  • CrossRef
  • Google Scholar
  • View reference in article

• 7

  ChoiJ.KimY.LeeS.LimH. (2025). Bridging social learning with technology: the use of a social robot in preschool development. Arch. Des. Res.38, 73–86. 10.15187/adr.2025.02.38.1.73

  • CrossRef
  • Google Scholar
  • View reference in article

• 8

  Criollo-CS.Guerrero-AriasA.Guaña-MoyaJ.SamalaA.Luján-MoraS. (2024). Towards sustainable education with the use of mobile augmented reality in early childhood and primary education: a systematic mapping. Sustainability16 (3), 1192. 10.3390/su16031192

  • CrossRef
  • Google Scholar
  • View reference in article

• 9

  DalimC. S. C.DeyA.PiumsomboonT.BillinghurstM.SunarS. (2016). “TeachAR: an interactive augmented reality tool for teaching basic English to non-native children,” in 2016 IEEE International Symposium on Mixed and Augmented Reality (ISMAR-Adjunct), (Merida: IEEE), 82–86. 10.1109/ISMAR-Adjunct.2016.0046

  • CrossRef
  • Google Scholar
  • View reference in article

• 10

  de HaasM.VogtP.KrahmerE. (2021). When preschoolers interact with an educational robot, does robot feedback influence engagement?Multimodal Technol. Interact.5 (12), 77. 10.3390/mti5120077

  • CrossRef
  • Google Scholar
  • View reference in article

• 11

  DignamC.SmithC. M.KellyA. L. (2025). The heart and art of robotics: from AI to artificial emotional intelligence in STEM education. J. Educ. Sci. Environ. Health11 (2), 151–169. 10.55549/jeseh.813

  • CrossRef
  • Google Scholar
  • View reference in article

• 12

  EryigitC. D.KüçükS.TasginA. (2025). Impact of augmented reality technology on geometry skills and motivation of preschool children. Educ. Inf. Technol.30 (4), 17225–17240. 10.1007/s10639-025-13631-4

  • CrossRef
  • Google Scholar
  • View reference in article

• 13

  FokidesE.SamiotiA. (2024). Teaching the seasons of the year to kindergarten students using desktop virtual reality: a comparative study. Qeios. 10.32388/IBDUBX

  • CrossRef
  • Google Scholar

• 14

  ForbesA.FalloonG.StevensonM.HatzigianniM.BowerM. (2021). An analysis of the nature of young students’ STEM learning in 3D technology-enhanced makerspaces. Early. Educ. Dev.32 (1), 172–187. 10.1080/10409289.2020.1781325

  • CrossRef
  • Google Scholar
  • View reference in article

• 15

  GarzónJ.AcevedoJ. (2019). Meta-analysis of the impact of augmented reality on students’ learning effectiveness. Educ. Res. Rev.27, 244–260. 10.1016/j.edurev.2019.04.001

  • CrossRef
  • Google Scholar
  • View reference in article

• 16

  GuyattG. H.OxmanA. D.VistG. E.KunzR.Falck-YtterY.Alonso-CoelloP.et al (2011). GRADE guidelines: 4. Rating the quality of evidence—study limitations (risk of bias). J. Clin. Epidemiol.64 (4), 407–415. 10.1016/j.jclinepi.2010.07.017

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 17

  HongQ. N.PluyeP.FàbreguesS.BartlettG.BoardmanF.CargoM.et al (2018). Mixed methods appraisal tool (MMAT), version 2018. Educ. Inf.34 (4), 285–291. 10.3233/EFI-180221

  • CrossRef
  • Google Scholar
  • View reference in article

• 18

  JiangH.ZhuD.ChughR.TurnbullD.JinW. (2025). Virtual reality and augmented reality-supported K–12 STEM learning: trends, advantages and challenges. Educ. Inf. Technol.30, 12827–12863. 10.1007/s10639-024-13210-z

  • CrossRef
  • Google Scholar
  • View reference in article

• 19

  JohnstonK.KervinL.WyethP. (2022). STEM, STEAM and makerspaces in early childhood: a scoping review. Sustainability14 (20), 13533. 10.3390/su142013533

  • CrossRef
  • Google Scholar
  • View reference in article

• 20

  Khan SoomroR. B.Basit SoomroA.Ali ShahS. T.MemonI. (2025). Integration of virtual reality and augmented reality into STEAM education: a meta-analysis. J. Adv. Res. Soc. Sci. Human.10 (9), 71–85. 10.61841/ameexk04

  • CrossRef
  • Google Scholar
  • View reference in article

• 21

  LampropoulosG. (2024). Teaching and learning natural sciences using augmented reality in preschool and primary education: a literature review. Adv. Mobile Learn. Educ. Res.4 (1), 1021–1037. 10.25082/AMLER.2024.01.013

  • CrossRef
  • Google Scholar
  • View reference in article

• 22

  LampropoulosG. (2025). Social robots in education: current trends and future challenges. Information16 (1), 29. 10.3390/info16010029

  • CrossRef
  • Google Scholar
  • View reference in article

• 23

  LeavyA.DickL.Meletiou-MavrotherisM.PaparistodemouE.StylianouE. (2023). The prevalence and use of emerging technologies in STEAM education: a systematic review of the literature. J. Comput. Assist. Learn.39, 1061–1082. 10.1111/jcal.12806

  • CrossRef
  • Google Scholar
  • View reference in article

• 24

  LeinonenT.VirnesM.HietalaI.BrinckJ. (2020). 3D Printing in the wild: adopting digital fabrication in elementary school education. Int. J. Art Des. Educ.39 (3), 600–615. 10.1111/jade.12310

  • CrossRef
  • Google Scholar
  • View reference in article

• 25

  MasturohU.IrayanaI.AdriliyanaF. N. (2024). Digitalization of play and games: artificial intelligence in early childhood education. TEMATIK J. Pemikiran Penelitian Pendidikan Anak Usia Dini10 (1), 21–30. 10.26858/tematik.v10i1.62195

  • CrossRef
  • Google Scholar
  • View reference in article

• 26

  NgA.KewalramaniS.KidmanG. (2022). Integrating and navigating STEAM (inSTEAM) in early childhood education: an integrative review and inSTEAM conceptual framework. EURASIA J. Math. Sci. Technol. Educ.18 (7), em2133. 10.29333/ejmste/12174

  • CrossRef
  • Google Scholar
  • View reference in article

• 27

  PageM. J.McKenzieJ. E.BossuytP. M.BoutronI.HoffmannT. C.MulrowC. D.et al (2021). The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ.372, n71. 10.1136/bmj.n160

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 28

  PanF.MelhemR. (2024). The effect of augmented reality storybooks on story comprehension and retelling in preschool children. Front. Psychol.15, 1459264. 10.3389/fpsyg.2024.1459264

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 29

  PanZ.LópezM.LiC.LiuM. (2021). Introducing augmented reality in early childhood literacy learning. Res. Learn. Technol.29, 10.25304/rlt.v29.2539

  • CrossRef
  • Google Scholar
  • View reference in article

• 30

  PelegA.LevyS. T. (2024). On the interrelationships between teachers, technologies, and children in makerspace environments: a comparative case study in kindergarten classrooms. J. Res. Child. Educ.38, 405–432. 10.1080/02568543.2023.2282662

  • CrossRef
  • Google Scholar
  • View reference in article

• 31

  PicardR. W. (1997). Affective Computing. Cambridge, MA: MIT Press.

  • Google Scholar
  • View reference in article

• 32

  RaptiS.SapounidisT.TselegkaridisS. (2024). Investigating educators’ and students’ perspectives on virtual reality enhanced teaching in preschool. Early. Child. Educ. J.53, 1107–1118. 10.1007/s10643-024-01659-z

  • CrossRef
  • Google Scholar
  • View reference in article

• 33

  Sánchez FernándezM. P.Hoyos RojasJ.Martínez TamayoL. (2025). Impact of educational robot R1 on the development of emotional intelligence in preschool children. GSC Adv. Res. Rev.24, 188–198. 10.30574/gscarr.2025.24.1.0188

  • CrossRef
  • Google Scholar
  • View reference in article

• 34

  SterneJ. A. C.SavovićJ.PageM. J.ElbersR. G.BlencoweN. S.BoutronI.et al (2019). Rob 2: a revised tool for assessing risk of bias in randomized trials. Br. Med. J.366, l4898. 10.1136/bmj.l4898

  • CrossRef
  • Google Scholar

• 35

  Videla-ReyesR.ArosM. (2025). Powerful ideas from a child cultivated in a FabLab. Pac. J. Technol. Enhanc. Learn.7 (2), 23–24. 10.24135/pjtel.v7i2.217

  • CrossRef
  • Google Scholar

• 36

  VistorteA. O. R.Deroncele-AcostaA.AyalaJ. L. M.BarrasaA.López-GraneroC.Martí-GonzálezM. (2024). Integrating artificial intelligence to assess emotions in learning environments: a systematic literature review. Front. Psychol.15, 1387089. 10.3389/fpsyg.2024.1387089

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 37

  WahyuningsihS.NurjanahN.RasmaniU.HafidahR.PudyaningtyasA.SyamsuddinM. (2020). STEAM Learning in early childhood education: a literature review. Int. J. Pedag. Teacher Educ.4, 33–44. 10.20961/ijpte.v4i1.39855

  • CrossRef
  • Google Scholar

• 38

  WangX.Abdul RahmanM. N. B.ShaharomM. S. N. (2024). The impacts of augmented reality technology integrated STEM preschooler module for teaching and learning activity on children in China. Cogent Educ.11 (1), 2343527. 10.1080/2331186X.2024.2343527

  • CrossRef
  • Google Scholar

• 39

  ZhangW.WangZ. (2021). Theory and practice of VR/AR in K–12 science education—a systematic review. Sustainability13 (22), 12646. 10.3390/su132212646

  • CrossRef
  • Google Scholar