Supplementary Figure 1. Plan and Research Design

In this case we applied the quasi-experimental pre-test/post-test control group design as shown at supplementary figure 1, a highly respected approach in educational technology research ( Creswell & Guetterman, 2019; Ary et al., 2020). The mixed-methods approach employed allowed for the fusion of quantitave data around learning gains and math anxiety with qualitative analysis focused on students’ perception of the experience in order to provide a nuanced description of effectiveness ( Johnson & Onwuegbuzie, 2020). This design also aligns with frameworks stating that interventions for math anxiety include an affective and cognitive measurement to assess the effects of digital game-based approaches ( Kahveci, 2022; Lagos San Martín, Ossa, Kaur Swaran Singh, & Mahmud, 2023). A diagrammatic representation of the mixed-method design incorporating quantitative and qualitative lines of inquiry is presented in Figure. The study starts with a pre-test through CUT and MARS-E to the two groups. After that, the intervention phase is implemented in which the experimental group uses an AR haptic game and control group is given traditional teaching. At the end of the intervention, both groups take a post-test with the same scales in order to measure changes and learning results. Results of the post-test measures are then examined to determine experiment effects, specifically within cognitive and affective domains. Simultaneously, the qualitative strand is performed to supplement results. This follows interviews and a teacher FGD to gain further insights into perceptions, experiences, guidance surrounding the interventions. The quantitative results and qualitative findings are then combined at the integrated findings stage, providing a more holistic understanding of the results, which connect numerical changes to experiential insights. This approach guarantees a nuanced analysis, which includes both how the intervention worked and what each study participant had experienced.

The sample consisted of 120 elementary school students (ages 9–11 years) drawn from two public schools in Yogyakarta. To ensure compliance with research ethics and educational regulations, the study was formally approved by both the research institution and the Department of Education and Youth Affairs, Yogyakarta City Government. Official research permission was granted, and the authorization letter is publicly accessible via its DOI link: https://doi.org/10.5281/zenodo.17120705 . Balanced representation was ensured by applying stratified random sampling ( Etikan & Bala, 2017). Half of the students joined one group: a.

Experiment group (N = 60): were taught mathematics using an AR and haptic feedback game.

All students shared the same prior exposure to mathematics according to baseline tests, consistent with sampling recommendations in experimental educational research ( Fraenkel, 2019). It is demonstrated in earlier research that math anxiety rates among children of this age typically associate with gender disparities, previous performance, and learning atmosphere ( Mitchell & George, 2022; Möhring, Moll, & Szubielska, 2024). This paper investigates the relationship between young Polish elementary schoolchildren’s self-reported mathematics anxiety and their actual numerical achievement. It was thus essential to make comparability of groups to minimize extrinsic bias.

The intervention was of six weeks duration, twice a week for 45 minutes. The game, “Bringing Numbers to Life”, was developed in Unity 3D, incorporating AR markers and haptic glove devices as an example of contemporary immersive learning design ( Radu, 2021; Santos et al., 2020). Its key features included: a.

AR Visualization – improving representation of fractions, geometry and multiplication ( Akçayır & Akçayır, 2022; Kaźmierczak et al., 2025).

Two weeks of training were provided to the teachers prior to implementation, which is aligned with guidance regarding fidelity when implementing technology-enhanced classrooms ( Bower, 2019). Further, it was found that the application of immersive AR and haptic experiences to tackle academic and emmotional barriers of learning (supported by sustainability-focused pedagogies for future ready education) such as those postulated by Mirza, Dutta, Tuli and Mantri (2025), were clearly conceptualized.

a.

Conceptual Understanding Test (CUT): A 25-item test Verified through review by experts (Cronbach’s α = .86).

The research proceeded in three phases: a.

Pre-test Phase: Both groups completed CUT and MARS-E.

Quantitative analysis was performed on SPSS 29. Group differences were analyzed using independent samples t-tests, paired t-test and ANCOVA as recommended in the best pratenice approach of educational experimental research ( Field, 2021). Modifier variables were control for and effect sizes (Cohen’s d) were computed to assess practical significance. MANOVA was performed to detect multidimensional changes for anxiety data. These analyses are consistent with suggestions that math anxiety be inspected in both cognitive and affective space ( Pizzie & Kraemer, 2023; Polydoros et al., 2025).

Thematic analysis of qualitative data was conducted using NVivo 14 according to Braun and Clarke’s (2019) six stages of thematic analysis. Themes were related to student engagement, motivation and connectedness, with an AR integration that have been emphasized in previous game-based learning and AR integration studies ( Lagos San Martín et al., 2023). The application of triangulation between test findings, observations, and interviews served to strengthen the reliability of results ( Lincoln & Guba, 2018).

This study was conducted in accordance with institutional ethical standards for research involving human participants ( BERA, 2018). Prior to data collection, ethical approval was obtained from the Research Ethics Committee of Yogyakarta State University with letter number: B/3018/UN34.13/TU.01/2025, and official permission was granted by the Yogyakarta City Education and Youth Office with letter number: 000.9/1581 ( Tanggu Mara, 2025b). As the children were in primary school (9–11 years), the written consent was sought from both the parents/guardians of students and school authorities. Moreover, the children also gave their verbal assent prior to participation and it was explained to the children in an age-appropriate manner that participation was voluntarily, they could withdraw from the study at any time. All recruited subjects were briefed on this, and told that these data will be kept confidential, anonymous, and for academic use only. Also, they were informed that in case of non-agreement and withdrawal at any time there would not be any prejudice. Verbal assent, as well as written parental consent, was utilised to verify understanding and comfort of the children participants in all aspects of the research ( Mitchell & George, 2022; O’Harra, 2022).

The prototype of AR and haptic feedback-based math game effectively combined three key components, which are the interactive AR objects, problem-solving missions within contexts embedded storylines, and the tactile response that emulates reality in contact with numbers. This set up allowed concepts from abstract math theory (place value, adding, subtracting and multiplying) to come alive as tangible interfacing. Evaluation of the prototype system by instructional designers and teachers indicated its compliance with usability, accessibility requirements, as well as its relationship to pedagogical aspects in primary classroom. These results are in a line with Garzón et al. (2020) and Ahmad (2020), whose reviews suggest that AR interventions improve engagement and results when underpinned by pedagogy. In addition, the provision of haptic feedback is a relatively new factor learning mathematics. Touching numbers with the game provided a bridge between abstract symbols and sensory input, re-emphasizing the kinesthetic learning pathways ( Crandall & Karadoğan, 2021; Hatira & Sarac, 2024). Consistent with observations of Marichal et al. (2023), haptic feedback was particularly beneficial for learners who needed multimodal input. This addresses previously professed standards in mathematics education research to create embodied experiences for children who struggle with abstract ways of thinking ( Nemirovsky, 2020; Bertrand, Sezer, & Namukasa, 2024).

The developed maths game app is available at https://kidar-playtime.lovable.app/ and serves as the central intervention of this research. The app combines Augmented Reality (AR) visualization with haptic feedback to create an augmented multisensory space for learners to interact with mathematical ideas. Before final release, the game was reviewed by domain experts Nsoccer systems, instructional design and maths who checked for content accuracy and its pedagogical integrity. Al-i-Imran (240), 50 Malay and nine hadith verses on head reduction of disease for six months. The resulting application was trialed with samples, which could prove the technical reliability and pedagogical impact. Comparable validation procedures have been documented in AR learning-related research to confirm the implementation fidelity ( Hanggara, Qohar, & Sukoriyanto, 2024; Pahmi et al., 2023).

Supplementary Figure 2. User Interface of AR Game

The user interface and interactive features of the AR-based haptic mathematics game are shown in supplementary figures 2. The design is centered on facilitating accessibility with simple and easy-to-understand controls, visual feedback, and tactile experiences to provide students exposure to “feeling” the basics of arithmetic. This multimodal approach not only promotes conceptual understanding, it also develops positive affective response, which is consistent with evidence suggesting that technology-rich learning environments can alleviate obstacles like math anxiety and promote engagement ( Wu et al., 2022; Fidan & Tuncel, 2019).

Table 1 Descriptive statistics of the research results for control and experimental groups. A trend of development was also found for the control group, even if very small (Mean = 58, s.d. = 10 pre-test; Mean = 62, s.d. = 9 post-test). In contrast, the experimental group demonstrated a marked improvement from pre-test (Mean = 59, SD = 11) to post-test (Mean = 82, SD = 8), indicating that treatment effects were large for participants in this condition.

Table 2 presents the statistical comparisons across conditions. The control group showed a small, non-significant improvement (Mean Difference = 4, p = .08). The experimental group showed a substantial gain from pre- to post-test (Mean Difference = 23, p < .001). Between-group comparison of post-test scores showed that the experimental group significantly outperformed the control group (Mean Difference = 20, p < .001), with a large effect size (Cohen’s d = 2.35). Effect sizes are reported only for between-group comparisons, consistent with APA standards.

Note. Cohen’s d is reported only for the between-group post-test comparison. The effect size is calculated using the standard formula: d = M exp − M ctrl SD pooled

with pooled standard deviation: SD pooled = SD exp 2 + SD ctrl 2 2

Using the post-test values: •

Experimental group: mean = 82, SD = 8

This effect size is very large and shows that the experimental group performed much better than the control group. Effect sizes are not shown for pre–post comparisons within each group because those calculations use a different method and cannot be directly compared to between-group effect sizes.

This result suggests the potential of combining AR and haptic feedback for improving not only students’ conceptual understanding, but also their motivation and involvement in learning mathematics ( Sáez-López et al., 2023). The AR tool was capable for making abstract mathematical conceptations, such as fractions and geometry, to be visualized in a 3D environment interactionally meanwhile the haptic feedback factor made tactile reinforcement who strenghtened the cognitive processes ( Shu et al., 2022). The post-test score of the experimental group not only indicates better achievement but also higher levels of attitudinal engagement, which may indicate lower math anxiety ( Ramirez et al., 2018; Erbaş & Demirer, 2025). In addition, qualitative feedback obtained in student interviews suggested that students enjoyed playing the game and found it motivating and less threatening than conventional approaches to learning. A number of students wrote that they “were no longer as afraid of making mistakes” and that “math became more fun when the numbers came alive.” This is in line with earlier research which have underscored that game-based learning employing immersive tech nologies can mitigate the pattern of negative emotions and the boost in academic selfconfidence in mathematics learning ( Fokides & Kapetangiorgi, 2022).

Table 3 displays the descriptive statistics of the scores on the CUT and MARS–E for experimental and control groups at pre and post tests. Conceptual understanding the two groups scored almost at a similar level in conceptual understanding at the beginning of the experiment with: Experimental group means score = 53.25, SD = 7.25; control group means score = 54.43, SD = 8.36 (Appendix b). Levels of mathematics anxiety were likewise approaching, with the experimental group having a mean of 66.48 (SD = 7.54) and for control group this was 66.70 (SD = 6.07). The fact that these numbers are comparable in both groups before the application of the intervention. Noticeable differences did arise afterward, however. The AR-haptic game experimental group showed highly significant improvement on conceptual understanding, with a mean of 73.22 (SD = 9.25), compared to the control group that used conventional methods (M1C¼62.43, SD1C¼9.42). The pattern was the same for mathematics anxiety. The experimental condition showed substantial improvements, with a post-test mean of 38.73 (SD = 8.76) as compared to the control’s mean 56.25 (SD = 8.32) little improvement was shown by the controls across assessments. Collectively, these findings suggest that the AR-haptic game intervention was effective in increasing students’ conceptual understanding and reducing their levels of MA compared to traditional teaching.

Supplementary Figure 3. Pre-Post CUT and MARS Treatment

Supplementary Figure 3 Explain there were statistically significant differences between their pre- and post-test results in the experimental group, compared with the control group. Students who were exposed to the AR-haptic game showed larger improvements in number sense, problem-solving accuracy and retention of arithmetic operations than students who did not play consequently. These results are consistent with Cai et al. (2022) and Wu et al. (2024) emphasized those AR enables deeper conceptual understanding by making visible abstract concept and enable manipulation to be part of representation. The tactile manipulations potentiated this effect, consistent with Pila et al. (2020) and Mutis and Oberemok (2025), who demonstrated that haptic-enabled tablets advanced preschoolers’ STEM learning, and underscored the taxonomical role of haptics in education. Moreover, Kaźmierczak et al. (2025) observed similar enhanced cognitive and non-cognitive effects, which assured our findings. Interestingly, the haptic aspect increased cognitive effect by involving multiple sensory modalities. Previous cognitive psychology studies show that learning under the condition of multi-sensory environment enhances information processing and memory retention ( Shams & Seitz, 2008; Zhang et al., 2023). In the present study, haptic reinforcement guaranteed learning was not only visual or auditory but rather embodied for durable and deeper knowledge of mathematical representations. This is particularly important in the case of younger students who stand to gain from a hands-on, experiential approach towards teach ( Montessori & Syafari, 2021).

One of the most surprising findings was that math anxiety decreased for the participants. The survey results revealed that students who had previously feared or avoided mathematics showed increased levels of confidence and fun after having played the AR-haptic game. They also referred to mathematics as “fun” and “less scary,” as they could mess with numbers in a playful setting. This is in agreement with Balt et al. (2022), Sammallahti et al. (2023) and Schoenherr, Schukajlow, Pekrun (2025), who argonued intervention promoting emotonal and cognitive coping together are superior in terms of reducing anxiety. AR was immersive and game-like, so it provided a safe environment to make mistakes learned from the failure as an exploration; similar to that of Kahveci (2022) and Lagos San Martín et al. (2023). Crucially, haptic feedback conveyed immediate tactile sensations that supported reassurance and confidence having been reported also in Yu et al. (2023) in which the AR had a positive effect on anxiety of learning physics via cognitive load reduction.

Psychologically, this illustrates the way in which embodied interactions can overcome affective hurdles. Mathematics anxiety has been associated with cognitive load and self-concept ( Ashcraft & Krause, 2007). Through redistribution of cognitive load across sensory channels and supportive tactile cues, the AR-haptic method reduced both the cognitive and emotional demands for problem-solving.

Supplementary Figure 4. Anxiety Changes Level After Treatment

Supplementary Figure 4 illustrates the relative increases or decreases in mathematics anxiety across control and experimental groups from pre-test to post-test. Only minimal decreases in math anxiety were found in the control group who reported a mean average score reduction from 3.1 to 2.9 on a Likert scale of agreement of one to five within session scores. This small decrease is likely small as it suggests that traditional methods of instruction were not able to adequately reduce negative affect related to mathematics learning. On the other hand, participants in the experimental group who played with an Augmented Reality (AR), Haptic Feedback mathematics game had a significant reduction on math anxiety from 3.2 to 1.8. This decline reflects that the users felt more engaging, less intimidating learning environment when using AR with multisensory haptic feedback. There was an increased “can do” attitude, enthusiasm and emotional positivity among students toward mathematical tasks. This result supports previous results that technology-based interventions can successfully reduce math anxiety by increasing engagement and reducing cognitive load ( Chang & Fang, 2020; Fidan & Tuncel, 2019). Furthermore, haptic feedback gave the learners a sense of touch which provided them with a multi-sensory experience ( Montgomerie, 2017) allowing teaching concepts in an abstract way but learning in a concrete manner and reduced affect barriers to learning ( Wu, 2021; Ramirez et al., 2018).

As can be seen in Table 4, there is a clear contrast between the two groups (control versus experimental) in terms of mathematics achievement and levels of math anxiety. Overtesting (F = 8.51, p = .**) and the control group evidence only minimal adiective gains from pre- to post-test (F = 4.90^ == > ^p =. 03, students in the experimental group (AR-Haptic math game) demonstrated a significantly higher level of achievement (F = 178.001, η ² = .14). This result implies that immersed MSM interaction by using AR and haptic may provide a significant advantage in promoting the conceptual understanding of students’ learning performance compared with that of traditional teaching.

With regard to math anxiety, the control group reported a small but significant decrease (F = 29.36, p < .001, η ² = .09). The experimental group, however, showed a greater and significant reduction in anxiety (F = 44.06, p < .001, η ² = .11). This showed that the typtactile approach in which multisensory and tactile engagement was integrated with visual learning, as compare to visuonly helped not only cognitive achievement but also emotional aspects making mathematics learning more favourable and fun environment.

Supplementary Figure 5. Result of Paired T-Test

Supplementary Figure 5 show the Paired T-Test radar chart sheds further light on this comparison, as it represents the results of paired samples t-tests performed for the experiments and controls at each measurement (CUT and MARS-E). Experimental group placement on the graph also demonstrates a significant increase in CUT scores, indicating much increased conceptual understanding post intervention. Meanwhile, the treatment group exhibits also a large change in MARS-E, as an overall decrease of scores reflects a significant decrease in math anxiety. The control group, on the other hand, plots significantly closer to the negative end of both CUT and MARS-E which demonstrates that although there were small increases in conceptual understanding and small decreases in anxiety, they were not as strong as those variables seen with the experimental group. Collectively, the visualization highlights the more pronounced combined impact of an AR-haptic game intervention on learning and anxiety in contrast to the relatively weaker effect from traditional teaching methods.

The double effect of improving cognitive benefits and decreasing affective obstacles, highlights the overall potential of AR and haptics in mathematics education. Previous research typically explored these two dimensions in isolation of one another, looking at either learning gains ( Ahmad & Junaini, 2020; Debrenti, 2024) or anxiety decrease ( Mitchell & George, 2022; Möhring et al., 2024). This manuscript shows that these two challenges could be solved in one shot through visual reinforcement with tactile cues and this is the main contribution of this work. By “making numbers real,” the intervention connected abstract knowledge with lived experience. By “alleviating anxiety,” it paved the way for students to approach mathematics learning emotionally and cognitively unfettered by fear. This pairing underscores an emergent pedagogical direction—designing multisensory and affect-aware educational technologies that is different from the traditional e-learning platforms ( Bulut & Ferri, 2023; Volioti et al., 2023). They also have implications for curriculum concordance. If AR-haptic systems are mainstreamed in classroom instruction, these could possibly function as not only supplementary but central pedagogical tools to transform mathematics learning and teaching ( Mirza et al., 2025; Zapata et al., 2024). This is especially essential in rural/underserved schools where in-service teacher knowledge about math anxiety may not be optimal and such technologies can mediate between affective support and cognitive scaffolding ( Putra, 2024; Ersozlu, 2024).

When interpreted within the limits of a quasi-experimental design, the results suggest that students in the experimental group showed greater improvement in conceptual understanding and reductions in mathematics anxiety compared to the control group. However, the findings indicate associations rather than causal effects, as random assignment was not feasible. The large effect size (d = 2.35) observed for post-test conceptual understanding should be interpreted with caution. Although some multimodal or game-based mathematics interventions with young learners have reported similarly strong effects, such values may also reflect measurement sensitivity, novelty of the technology, or heightened engagement rather than the intrinsic impact of AR–haptic components alone.

The qualitative themes help illuminate the processes through which students may have experienced the learning environment. “Embodied confidence” and “cognitive anchoring” are consistent with embodied cognition literature, which posits that sensorimotor interaction can support conceptual grounding and reduce cognitive load during abstract reasoning. The theme of “situated enjoyment” suggests that affective benefits may have contributed to the reduced anxiety scores observed quantitatively. Nonetheless, these interpretations remain exploratory and cannot determine directionality or causal mechanisms.

Embodied cognition models argue that mathematical reasoning is not solely symbolic but arises from the coordination of perception and action ( Abrahamson et al., 2020; Nemirovsky & Van Rooij, 2010). In this view, sensorimotor engagement—such as touching, manipulating, or bodily simulating mathematical structures—creates material anchors for conceptual understanding.

The AR–haptic system used in this study aligns with this mechanism in two ways: 1.

AR visualization structures perceptual attention.

By presenting spatially meaningful representations (e.g., fraction partitions, place-value blocks), AR enables learners to see the underlying mathematical structure rather than work exclusively with abstract symbols. 2.

Haptic feedback recruits bodily enactment.

The tactile response invites learners to act on numbers through gesture-like interactions, shaping stable sensorimotor patterns. This resonates with Abrahamson et al.’s principle that conceptual insight emerges when learners coordinate perceptual cues with physically meaningful actions.

The strong pre–post learning gain in the experimental group may therefore reflect a shift from purely symbolic processing to sensorimotor grounding. Importantly, this theoretical lens provides a more nuanced explanation than attributing gains to engagement alone.

Several plausible confounders may have influenced outcomes: •

Novelty effect: The excitement of using AR and haptic devices may have temporarily elevated motivation and engagement.

These variables constrain the extent to which observed gains can be attributed specifically to the AR–haptic learning environment.

Because the study implemented an integrated AR–haptic system, the design does not isolate the unique contributions of each modality. It remains unclear whether improved outcomes stem from: •

the visuospatial affordances of AR,

Future studies should employ dismantling designs (e.g., AR-only vs. haptic-only vs. AR–haptic combined) or mediation models to test which components contribute most strongly to cognitive and affective outcomes.

The findings provide a promising proof of concept for the feasibility of implementing multisensory, embodied learning tools in primary mathematics classrooms. While the study does not establish theoretical innovation or causal effects, it demonstrates how embodied cognition principles can be operationalized using accessible AR–haptic technologies. Future research should consider longitudinal designs to examine retention, follow-up measures to evaluate sustained anxiety reduction, and randomized trials to strengthen inferences.

The results obtained in Table 5 reveal that the AR-Haptic mathematics game was an overall positive learning experience for students. The majority of participants stated that the game was engaging and pleasant to look at, which increased capacity for capturing attention as well as motivation. Furthermore, the augmented reality combined with haptic feedback supported meaningful understanding through visual elaboration of mathematical concepts which made overall abstract ideas easier to conceptualize. The students also liked the game’s simplicity and its ability to be self-paced so that they could actually go back to play again and reinforce what they had learned. However, some limitations were noticed in this study. Some students had longer trace and response time, and some visual elements were not always easy to follow, which could disturb the interaction conceptual understanding. In order to address such problems, participants had proposed issues like reducing tracking delay time, enhancing visual elements (i.e. fonts, objects) as well as adding additional interactive features to increase attractiveness and usability offer participants of this study all held different opinions on web3D usability according to their previous experience with web3D transactions and activities. The study shows however that AR combined with haptic feedback has distinctive cognitive and affective effects on mathematics learning for elementary school children. The prototype game designed in this study delivered rich and engaging multisensory experiences that instantly enhanced how well students grasped the concepts being studied, their problem-solving skills even remediate math-anxiety an established roadblock to performance in early education.

The results also support the claim that, in learning environments that facilitate multisensory interaction, students are able to connect with abstract mathematical concepts and make them relevant by grounding them on a tangible foundation, rendering the working process exciting and meaningful. In addition, the affective results show that AR-haptic systems have promise in providing emotionally supportive and motivating materials for students to engage with mathematics without fear. Thus, the work emphasises technology’s two-fold function as not only a mechanism to improve academic achievement but also as an emotionally and psychologically supportive education tool.

This study does not attempt to extend embodied cognition theory. Instead, it provides a proof of concept showing how existing theoretical principles—such as sensorimotor grounding, perceptual coupling, and multimodal learning—can be instantiated through AR–haptic technologies in elementary mathematics classrooms. The results affirm, rather than expand, embodied cognition frameworks by demonstrating their practical realizability and educational impact ( Acevedo-Charry et al., 2021; Garzón-Duque et al., 2021; Cai et al., 2022). Augmented Reality for Education: A Multimethod Literature Review but hardly two at the same time. By integrating AR visualization with haptic feedback, the current work shows how learning technologies can scaffold both cognitive success and overcome affective resistance. This is consistent with, and extends, constructivist and embodied cognition perspectives which propose that mathematics learning is most effective when abstract reasoning has roots in real sensory experience ( Nemirovsky & Van Rooij, 2010; Zhang et al., 2023). Embodied cognition frameworks argue that mathematical reasoning arises from perceptual–motor activity, including gesture, spatial manipulation, and tactile interaction. In this perspective, AR affords dynamic visuospatial representations while haptic feedback provides sensorimotor grounding of mathematical structures. This theoretical linkage informed the game’s design, enabling learners to act on numerical relationships rather than treat them as purely symbolic abstractions ( Nemirovsky & Van Rooij, 2010; Zhang et al., 2023). In addition, the finding that affective states directly impact academic performance further supports findings in educational psychology of emotions constituting an integral part, rather than a mere addendum, of learning ( Pekrun, 2021).

The contributions of this research are as follows: a.

For Teachers and Schools: Mathematics of AR-haptic games can lay out the resources for teachers, use augmented reality (AR) to equip teachers with tools that support their students’ academic and emotions. In the context of limited teaching resources, such tools offer personalized support when teacher support is not available or adequate.

Despite the promising findings, several important limitations must be acknowledged beyond sample size and intervention duration. First, the technological constraints of the AR–haptic system were not examined systematically. Hardware-related issues—such as marker tracking delays, calibration inconsistencies, battery performance, and the precision of the haptic glove’s feedback—may have influenced the smoothness of interaction and consequently shaped students’ learning experiences. The absence of metrics on device accuracy and reliability limits the ability to determine how much the observed outcomes depend on technology fidelity rather than pedagogical design. Second, the study did not evaluate the pedagogical transferability of the intervention. Although teachers received two weeks of training, the required level of technological familiarity, classroom orchestration skills, and troubleshooting competence may challenge broader implementation. Classroom management dynamics—especially in large or resource-limited settings—were not assessed, making it unclear how feasible it would be for typical teachers to integrate such tools into everyday mathematics instruction without substantial support. Third, because AR and haptic feedback were integrated into a single system, the study cannot determine the specific contribution of each modality. Visual augmentation, tactile cues, and their combination may produce different effects on learning and anxiety. Without comparative conditions, the results cannot disentangle which component—or interaction of components—played the most meaningful role.

Future research should pursue several important directions. Longitudinal studies are needed to examine whether the observed cognitive and affective benefits persist over time, particularly with respect to conceptual retention and the stability of anxiety reduction. Cost–benefit analyses would help determine the practicality and scalability of AR–haptic systems in diverse school environments, including assessments of device maintenance, training requirements, and classroom integration effort. Comparative or dismantling designs—such as AR-only, haptic-only, and combined AR–haptic conditions—would allow researchers to isolate and evaluate the specific contributions of each modality. In addition, future work should investigate the technological reliability of hardware components through systematic calibration assessments, error-logging, and analyses of feedback precision. Finally, research should explore pedagogical transferability by examining how teachers with varying levels of expertise adopt, adapt, and sustain the use of multisensory technologies in real classrooms. Such studies could include professional development models, teacher beliefs about embodied learning, and organizational factors influencing adoption. These lines of inquiry would strengthen understanding of both the practical and theoretical dimensions of integrating embodied, multisensory learning technologies into elementary mathematics education.

At last, the study is another piece of evidence to the emerging technology-enhanced learning which may be cognitively empowering and affective supportive as well. In bringing together AR, haptic feedback and mathematics education this research provides a route for creating learning environments that are both more successful as learning spaces but also more humane. With educators globally looking for innovative ways to enhance STEM educational outcomes, AR-haptic games offer a potential pathway towards both capability and confidence building among children.