Ethnomathematics-Based Augmented Reality and Students’ Numeracy Literacy: A Quasi-Experimental Study Across Levels of Mathematics Self-Efficacy
Journal of Information Technology Education: Research · 2026 · 25 · 44
DOI: 10.28945/5902 ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal of Information Technology Education: Research, 2026. Verfasst von Eviyana Eviyana und 4 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY-NC 4.0).
Abstract (Originalsprache)
Aim/Purpose: This study examines three questions about the classroom use of ethno-mathematics-based Augmented Reality (AR) in secondary mathematics: whether students who were taught geometric transformation with a purpose-built AR application, GeoTrans AR Souraja, differ from students who were taught the same content conventionally in geometric numeracy literacy once prior attainment is taken into account; whether numeracy literacy differs across levels of mathematics self-efficacy; and whether self-efficacy moderates the effect of the instructional medium. Self-efficacy is treated throughout as a pre-existing moderator rather than a mediator.
Background
Indonesian students’ literacy and numeracy performance both remain low: PISA 2022 placed Indonesia 68th of 81 participating systems in mathematics, with a mean score of 366 against an OECD average of 472, and similarly low in reading, with a mean score of 359 against an OECD average of 476, and the 2024 national education report indicates that only 67.14% of senior secondary students reach minimum numeracy competency and 72.23% reach minimum literacy competency. Geometric transformation is a recognised bottleneck because it requires students to coordinate visual representation with symbolic and coordinate-based reasoning. Augmented reality offers dynamic three-dimensional visualisation of these transformations, and ethnomathematics provides a culturally situated context in which the transformations already exist as physical structures. However, whether the benefits of such media are distributed evenly among learners with differing self-efficacy remains largely untested. Methodology: A quantitative quasi-experimental study using a non-equivalent control-group design analysed within a 3×2 factorial framework (three self-efficacy levels × two instructional media). Participants were 61 eleventh-grade students in two intact classes at a single private senior high school in Palu, Central Sulawesi, purposively selected on the basis of comparable prior-semester achievement; one intact class (n = 31) received instruction using the ethnomathematics-based AR medium, and the other (n = 30) received conventional instruction. Both classes were taught the same content over two 45-minute meetings each (4 teaching hours per class) using the problem-based learning model, so that the conditions differed in medium rather than in pedagogical model. Numeracy literacy was measured with an open-ended geometric transformation task comprising four scored sub-items. Mathematics self-efficacy was measured with a 37-item four-point Likert questionnaire covering magnitude, strength, and generality (α = 0.863), analysed both as a continuous variable and, for the factorial model, categorised at the class mean ± 1 SD, the split specified in the original analysis plan, with a whole-sample split reported as a sensitivity check. Analyses comprised descriptive statistics, baseline comparison, analysis of covariance with pretest as covariate, two-way ANOVA, and regression with continuous moderation, with effect sizes and 95% confidence intervals reported throughout. Contribution: The study makes two distinguishable contributions. As an instructional contribution, it documents the design and classroom implementation of GeoTrans AR Souraja, a marker-based Android application in which four architectural elements of the Souraja traditional house of the Kaili people are mapped onto translation, reflection, rotation, and dilation. As an empirical contribution, it provides preliminary quasi-experimental evidence from a single school on the association between such a medium and geometric numeracy literacy, along with a transparent account of the statistical limits of that evidence. The strength of the empirical claim should be judged separately from the value of the instructional artefact.
Findings
The two intact classes were not significantly different at baseline (experimental M = 52.82, SD = 25.35; control M = 45.00, SD = 23.58; t(59) = 1.25, p = 0.216), although the experimental class held a non-trivial initial advantage of 7.82 points, d = 0.32. On the posttest, students in the AR condition scored higher (M = 77.68, SD = 14.51) than students in the conventional condition (M = 69.47, SD = 17.01). The unadjusted difference of 8.21 points was statistically significant but marginal in the two- group comparison, t(59) = 2.03, p = 0.047, 95% CI [0.12, 16.30], Hedges’ g = 0.51, and remained significant when self-efficacy was added as a second factor, F(1, 55) = 5.60, p = 0.021, partial η² = 0.09. It did not survive adjustment for prior attainment: in analysis of covariance, the covariate was a substantial predictor, F(1, 58) = 10.18, p = 0.002, partial η² = 0.15, while the adjusted difference fell to 6.27 points, 95% CI [−1.36, 13.90], F(1, 58) = 2.70, p = 0.106, partial η² = 0.04. Rank-based robustness checks agreed (Quade ANCOVA p = 0.177; Mann–Whitney p = 0.087). Neither the main effect of self-efficacy, F(2, 55) = 2.37, p = 0.103, partial η² = 0.08, nor the medium × self-efficacy interaction, F(2, 55) = 1.87, p = 0.165, partial η² = 0.06, reached significance, and both were also null when self-efficacy was modelled as a continuous score (b = 0.18, p = 0.403 and b = 0.19, p = 0.657; zero-order r = 0.196, p = 0.131). A sensitivity analysis showed the design could detect only large effects (f ≥ 0.37 for the main effect, f ≥ 0.41 for the interaction), so all null results are inconclusive rather than evidence of equivalence. All five instructional effectiveness indicators were rated good or better, although classical mastery reached 77.4%, only narrowly above the 75% criterion. Recommendations for Practitioners: Teachers may reasonably trial ethnomathematics-based AR as one option for making geometric transformation concrete, particularly where a local cultural artefact already embodies the transformations to be taught. The present evidence supports trialing the approach rather than adopting it as an established improvement, and schools should weigh the required access to devices, marker printing, and teacher preparation against a modest, imprecisely estimated benefit. Recommendation for Researchers: Replications should use multiple schools with randomised or carefully matched assignment, samples large enough to support factorial tests of aptitude–treatment interaction, and pretest-adjusted analysis reported as standard. Self-efficacy should preferably be modelled as a continuous predictor rather than split into categories. Component-controlled designs are needed to separate the contribution of three-dimensional visualisation from that of cultural context, novel software, and novelty itself. Impact on Society: Where it proves effective, mathematics instruction anchored in regional cultural heritage may support numeracy learning while giving students a reason to engage with heritage at risk of erosion. The Souraja house is used here as a worked example of how a locally significant structure can serve as a mathematical object of study rather than as decoration. Future Research: Priorities include longer interventions with delayed retention testing, extension to other Indonesian cultural contexts and educational levels, usability evaluation of the application, and qualitative process evidence on how students reason while manipulating AR representations.
Zitieren
APA
Eviyana, E., Ismaimuza, D., Sudarsana, I. W., Sukayasa, S., & Pathuddin, P. (2026). Ethnomathematics-Based Augmented Reality and Students’ Numeracy Literacy: A Quasi-Experimental Study Across Levels of Mathematics Self-Efficacy. Journal of Information Technology Education: Research, 25, 44. https://doi.org/10.28945/5902
Vancouver
Eviyana E, Ismaimuza D, Sudarsana IW, Sukayasa S, Pathuddin P. Ethnomathematics-Based Augmented Reality and Students’ Numeracy Literacy: A Quasi-Experimental Study Across Levels of Mathematics Self-Efficacy. Journal of Information Technology Education: Research. 2026;25:44. doi:10.28945/5902
BibTeX
@article{Eviyana2026,
author = {Eviyana, Eviyana and Ismaimuza, Dasa and Sudarsana, I Wayan and Sukayasa, Sukayasa and Pathuddin, Pathuddin},
title = {Ethnomathematics-Based Augmented Reality and Students’ Numeracy Literacy: A Quasi-Experimental Study Across Levels of Mathematics Self-Efficacy},
journal = {Journal of Information Technology Education: Research},
year = {2026},
volume = {25},
pages = {44},
doi = {10.28945/5902}
}
EUVIMED