Improving Reading Ability Using Augmented Reality

Abstract

This study investigates the integration of learning theories with personalized technologies, focusing on the use of Augmented Reality (AR) in reading instruction. Its primary aim is to conduct a systematic literature review of research employing AR to support readers with the complexities of the reading process. The review focuses on literacy development in children from kindergarten through age twelve, encompassing both typically developing readers and those with reading difficulties. It is suggested that AR can contribute to inclusive education by offering adaptable and engaging learning experiences that meet diverse learner needs. Guided by clearly defined inclusion and exclusion criteria, the study analyzes key elements of research design, the types of AR technologies utilized, and the educational outcomes reported. Furthermore, it critically examines the limitations of the current body of evidence in this field.

1. Introduction

Proficiency in reading is essential for acquiring knowledge, social engagement and further well-being (Hulme & Snowling, 2012). Early difficulties in reading can have profound negative effects, particularly during early primary education accumulating ‘Matthew effects’ (i.e., inequalities lead to cumulative disadvantages via negative feedback) (Stanovich, 1991). Low levels of reading comprehension contribute to increased school dropout rates and limited access to future opportunities regarding career and further education (Alexander et al., 2001).

Digital technology contributes to reading education offering innovative practices that promote literacy skills, especially for students who face reading challenges. Several studies highlight the effectiveness of digital tools in teaching, fostering motivation and participation of students while providing a supportive learning environment. For example, Fälth and Selenius (2024) explored the perspectives of primary school teachers toward the use of digital technology in early reading and writing instruction. The participants expressed positive attitudes, emphasizing the crucial role of technology in engaging diverse learners in general classroom. Specifically, teachers highlighted the contribution of digital technology in personalized learning experiences related to the different learning needs.

Specific digital tools have been shown to provide targeted support in literacy by including text-to-speech software, interactive e-books, and augmented reality applications (Alanazi & Abdulkader, 2024; Paudel & Acharya, 2024; Zhao et al., 2025). MacArthur et al. (2001) highlight that text-to-speech software enhances comprehension and engagement by reading text aloud, making it particularly beneficial for students with learning disabilities. Similarly, Korat and Shamir (2008) found that interactive e-books with animations, sounds, and clickable words enhance vocabulary, comprehension, motivation, and retention of young learners. Audiobooks play a significant role in developing reading fluency and comprehension, as they help young learners to enrich vocabulary and internalize correct pronunciation and intonation through listening (e.g., Dalton & Proctor, 2007; Joshi, 2006). Additionally, reading apps and games promote engagement and reading development as they use interactive and gamified elements and provide immediate feedback (Zheng et al., 2020). Schmitt et al. (2018) explored the impact of web-based educational games on early literacy development (phonological awareness, letter recognition, and vocabulary acquisition) in preschool-aged children and highlighted the importance of content quality, frequency of use, and guided play. Speech recognition tools assist students with learning disabilities by enabling verbal practice and converting speech to text, helping them overcome writing challenges (e.g., Bouck & Flanagan, 2010).

1.1. Augmented Reality (AR) and Reading Instruction

Furthermore, Augmented Reality (AR) and its various applications (e-texts, mobile applications, interactive AR books, AR games) support knowledge acquisition and literacy development, by making texts more interactive and engaging (H. Y. Chang et al., 2022; Y. Liu et al., 2023; S. Liu et al., 2024a; Tobar-Muñoz et al., 2017). Defined as the overlay of 3D virtual content (images, audio, video) onto the physical world (Azuma, 1997), AR has evolved from marker-based systems to advanced immersive technologies. Studies grounded in constructivist theory frequently use experimental designs to assess AR’s educational benefits (Wojciechowski & Cellary, 2013). Key outcomes include improved reading motivation, attention, comprehension, storytelling skills, and overall academic performance (Danaei et al., 2020; Fernández-Batanero et al., 2022; Y. Liu et al., 2023; S. Liu et al., 2024a).

The development of AR applications enhances traditional learning environments (i.e., printed materials for books) with interactive features such as audio, 3D visual effects, and animations and create dynamic and immersive learning experiences, enabling meaningful navigations between physical and digital content (Akçayır & Akçayır, 2017; Besa, 2021; Fernández-Batanero et al., 2022; Wu et al., 2013; Wang et al., 2014).

When AR is implemented during instruction, it can benefit students by developing motivation, communication, and interaction (Iatraki & Mikropoulos, 2025). Students achieve deeper learning when multiple methods of representation, engagement, and expression are provided. In addition, AR applications facilitate the implementation of evidence-based reading strategies such as video modeling, repetition, segmentation, blending, and self-questioning (Howorth et al., 2019). In AR learning environments, students may have the sense of presence by feeling of ‘being there’ and may interact with the environment in real time through multisensory channels in a way that approximates reality, depending on the level of immersion (Slater & Sanchez-Vives, 2016). Thus, it has attracted the research interest of technology orientated research teams seeking to explore possible mediums to improve performance and quality of life of individuals who most need it.

Chen and Tsai (2012) found that an interactive AR system enhanced reading engagement and information retrieval skills among elementary students. Although AR glasses and headsets have the potential for immersive reading experiences, their current use in education remains limited due to high costs and accessibility challenges (Radu, 2014). Ibáñez et al. (2014) demonstrated that AR mobile apps enriched with contextual visuals significantly increased motivation and comprehension in primary school students. Similarly, Wu et al. (2013) developed a tablet-based AR storytelling system that combined text, audio narration, and 3D graphics to support children with reading difficulties, resulting in improved fluency, vocabulary, and interest in reading. Quintero et al. (2019) conducted a systematic review analyzing the role of AR in promoting educational inclusion over the previous decade. The findings revealed that AR contributed positively to inclusive education by enhancing student motivation, engagement, and interaction, especially among learners who faced difficulties with traditional instructional methods.

In their systematic review, Garzón et al. (2020) further highlighted the benefits of AR delivered through tablets and smartphones, emphasizing its ability to create multisensory experiences that boost motivation and understanding. The results showed consistently positive effects of AR on learning outcomes, especially as regards motivation, engagement, and conceptual understanding, with a moderate to high overall effect size. While the findings support the integration of AR into classroom practices, the authors also noted common methodological limitations in the reviewed studies, such as short intervention durations, lack of control groups, and limited attention to accessibility. The study concludes by recommending more longitudinal and inclusive research designs to better assess the sustained and equitable impact of AR in K–12 settings.

Similar conclusions were drawn in the systematic review of Yenioglu et al. (2021), which examined the use of AR on students with learning disabilities. The interactive, multisensory, and personalized nature of AR proved especially beneficial for learners who face challenges with traditional text-based instruction. Similarly, Paudel and Acharya (2024) reviewed 29 studies (2015–2023) enrolling children with dyslexia and concluded that AR emerged as a leading intervention due to its interactive, multisensory features that support phonological decoding and real-time feedback.

1.2. The Contribution of AR in Reading Development from a Pedagogical Perspective

There is ongoing debate about its benefits and drawbacks of AR use for reading development. Research suggests that AR may enhance reading performance and motivation. Its interactive and multimodal features such as the integration of verbal and non-verbal elements, and the presentation of content through engaging formats like images, stories, and text, contribute to these positive effects. These features can facilitate connections between prior knowledge and new information, strengthen cognitive skills, and support the overall development of reading abilities (Danaei et al., 2020). Different views, however, highlight the risk of cognitive overload and the potentially complex and distracting nature of AR technology, which may hinder rather than support reading development.

The benefits of implementing AR have been well documented in several studies. In the context of literacy development, advantages involve reduced cognitive load, enhanced collaborative learning, increased engagement and interest, a more positive and entertaining learning experience, improved information management, greater ability to complete reading tasks, as well as support for problem-solving and creativity. However, there are also significant drawbacks. These include high costs, technical and pedagogical challenges, student frustration and distraction, limited availability of free AR applications, and a lack of professional development and support for teachers (Çetin, 2022). Despite these challenges, AR holds potential for meaningful implementation in reading instruction. While its limitations should not be overlooked, its pedagogical benefits offer compelling reasons to consider its use for children with or without reading difficulties. However, there is still limited research on the impact of AR environments and their innovations in special education, especially regarding their applicability to reading instruction for students with and without reading difficulties.

The pedagogical implications of this endeavor must be acknowledged in an effort to understand its merits. Based on the design principles summarized by Y. K. Chang et al. (2024), the major theories that inform pedagogy and instructional design and explain the impact of AR on reading development may be summarized as follows.

Theories involving cognitive processing, such as Mayer’s (1997, 2002, 2009) Multimedia Learning Theory, emphasize principles like coherence, modality, and redundancy, which are essential when designing AR experiences. These principles are grounded in Cognitive Load Theory (Sweller, 1988) and Dual Coding Theory (Paivio, 1971, 1986), which highlight the importance of managing mental effort and integrating verbal and visual information, respectively, for effective learning. Mayer’s Multimedia Learning Theory (Mayer, 2002, 2005) is firmly grounded in attention theories (e.g., Broadbent, 1958; Treisman, 1964), which emphasize how learners filter and focus on relevant information. This is particularly relevant in AR environments, where a variety of modalities such as highlighting, directional arrows, audio cues, and animations can be used strategically to guide attention and reduce cognitive distraction.

In addition, embodied cognition (e.g., Barsalou, 1999, 2008; Wilson, 2002) is closely linked to immersive technologies, which use embodied experiences, such as manipulating virtual objects to foster deeper conceptual understanding. In school settings, simpler forms involve kinesthetic learning strategies, allowing students to connect abstract concepts with physical experience. Embodied cognition along with immediate interaction and feedback, gamified or situated activities, provide an ‘inside’ experience with deep concentration even with challenging activities. This optimal experience has been recognized with the term ‘flow’ (Csikszentmihalyi, 1990) and is meant to be rewarding, increasing intrinsic motivation and adherence to the task.

Furthermore, constructivist approaches to learning emphasize the active construction of knowledge—often supported through strategies like scaffolding—and align well with AR-based learning environments. These approaches include Experiential Learning (Kolb, 1984), Discovery Learning (Bruner, 1961), and Situated Learning (Lave & Wenger, 1991), all of which stress the importance of context, exploration, and real-world engagement in the learning process.

Moreover, principles from developmental psychology inform task design (Radu & MacIntyre, 2012). AR features such as interaction type, language and narration, scaffolding, visual stimuli, and feedback, should align with age-appropriate developmental levels to ensure that learners are neither overwhelmed nor under-challenged. This congruence helps maximize engagement, content comprehension, and learning outcomes. Effective AR design should align with pedagogical frameworks based on learning theories and evidence-based practices considering learners’ individualized needs.

2. Methods

According to the Universal Design for Learning (UDL), which promotes inclusive and flexible learning environments (CAST, 2018), the present review focuses on the implementation of AR in supporting reading development among primary school students with and without reading difficulties. UDL, originally developed by CAST in the 1990s and grounded in research on cognitive neuroscience, provides a framework for anticipating learner variability and proactively designing instruction that minimizes barriers and maximizes accessibility. The framework is structured around three guidelines: multiple means of engagement (motivating and sustaining interest in learning), multiple means of representation (providing information in different formats), and multiple means of action and expression (offering learners diverse ways to demonstrate understanding). One of the main goals of this review is to identify the pedagogical features of AR interventions. Not UDL was selected as the analytic lens to evaluate how AR supports literacy development for diverse learners. In particular, this framework enables systematic examination of whether and how AR technologies provide engaging, multimodal, and flexible learning opportunities that are inclusive of children who experience reading difficulties.

The key aspects of the study concern the investigation of research designs, the types of AR technologies employed, and the outcomes achieved. Additionally, the limitations of the existing evidence in this area are discussed. The research questions were as follows:

• What augmented reality (AR) reading interventions have been employed to support primary school students’ reading skills?
• What are the strengths and weaknesses of research studies that used AR to help children improve their reading skills, according to the studies that were reviewed?
• What are the limitations of these studies that may introduce threats to their validity?

2.1. Data Sources/Search Strategy

Articles examining the impact of AR on primary school students with reading difficulties were identified using a Boolean combination of keywords related to “augmented reality” and “reading difficulties”. The electronic academic databases were: Google Scholar (n = 246), Scopus (n = 38), ERIC (n = 66) and ScienceDirect (n = 328) and the publisher’ websites: SAGE (n = 79) and Taylor & Francis (n = 36). These sources were selected due to their high-quality, peer-reviewed research in the field of education. The final search string was (“augmented reality” or AR) AND (instruction OR teaching OR intervention) AND (reading OR comprehension OR “reading difficulties” OR dyslexia OR “reading problems”) AND primary. A total of 793 records were retrieved from peer-reviewed journals and conference proceedings published in English between 2014 and 2024, using the available filters provided by each database.

All identified records were imported into the EndNote™ (version 25) reference management software. Using EndNote’s duplicate detection feature, we automatically removed duplicate entries, resulting in the unique records for further screening. The procedure for obtaining the included studies was divided into the stages of identification, screening, and inclusion, as presented in Figure 1. Both researchers independently screened titles and abstracts for relevance, then assessed full-text articles based on the inclusion criteria. During screening, non-relevant studies (studies integrated other digital tools, non-empirical studies, such as assessment studies or user experience studies, or studies enrolling students with other special educational needs) were excluded. Moreover, reviews, meta-analyses, dissertations, theoretical studies and reports were also excluded. Finally, based on the rigorous selection process, a total of 11 studies were included in the review.

2.2. Study Selection/Eligibility Criteria

The studies met the following inclusion criteria: (i) publication dates should be in the specified scope of the last decade of 2014–2024 published in a peer-reviewed journal or conference, (ii) studies should be written in English language, (iii) studies should involve participation of students with or without reading difficulties in primary education (up to age 12), (iv) interventions using AR should aim to improve reading performance (decoding and/or comprehension) in first language, and (v) studies should clearly report results of intervention effectiveness (e.g., before and after intervention).

2.3. Data Extraction and Coding

A structured data extraction form was developed to systematically collect relevant information from the included studies. The extracted data were organized into three main categories: publication metadata, AR features, and study characteristics. Publication metadata included the publication year, author names, and source (i.e., journal or conference). Study characteristics included detailed information about the participants (e.g., number, age, gender, type of disability), the instructional setting and interventionists involved, the number of sessions delivered, the content and the learning objectives. The relationship between independent variables (i.e., interventions) and dependent variables (i.e., learning outcomes) was also documented. Data necessary to evaluate methodological rigor were also collected, in alignment with the Quality Indicators for Research in Special Education (Brantlinger et al., 2005; Gersten et al., 2005; Horner et al., 2005; Thompson et al., 2005).

For AR features, the analysis focused on their pedagogical added value using the Universal Design for Learning (UDL) framework (CAST, 2018). In line with the second research question, UDL was selected as the analytic lens to evaluate how AR interventions were designed to address learner variability in reading development. Data were therefore coded according to the three UDL guidelines: (1) multiple means of engagement (e.g., gamification, interactive prompts, adaptive feedback), (2) multiple means of representation (e.g., animated visuals, synchronized narration, multimodal scaffolds), and (3) multiple means of action and expression (e.g., verbal responses, marker-based interactions, collaborative dialog). Additionally, technological characteristics of AR applications (e.g., marker-based systems, mobile compatibility, visual and auditory stimuli) were documented to examine how design features intersected with UDL principles. This allowed the review not only to map the technical affordances of AR, but also to assess their pedagogical relevance for inclusive and flexible literacy instruction.

3. Results

The results and discussion are organized based on the characteristics of the studies, the participants and settings, and the research questions addressed by the empirical studies included in this review.

3.1. What Augmented Reality (AR) Reading Interventions Have Been Employed to Support Primary School Students’ Reading Skills?

Table 1 provides an aggregated overview of the included articles, outlining key study characteristics, AR technological features, and pedagogical approaches. All the studies focused on academic achievement, emphasizing reading and reading comprehension. Eight of the eleven included articles were published after 2020, indicating a growing trend in the use of AR reading interventions in primary education. Six studies employed in this review involved research on AR interventions regarding reading in primary education and mostly focused on students without reading difficulties. The rest five studies targeted children with reading difficulties perhaps due to Attention Deficit Hyperactivity Disorder (ADHD) (Lin et al., 2016) or due to ADHD and comorbidities (Tosto et al., 2020) and students with reading difficulties (Ok et al., 2021) or low achievers (Nasongkhla & Sujiva, 2023; Poobrasert & Luxameevanich, 2023). Most of the studies provided detailed and replicable descriptions of their participants. The two single-subject studies included two to three participants each, while the group studies involved between 14 and 120 participants. Across the 11 documents reviewed, a total of 540 participants were represented.

The review of the studies reveals a consistent pattern of significant gains across various literacy-related outcomes, such as reading, word recognition and reading comprehension (Table 1). The studies that involved experimental groups demonstrated higher levels of reading comprehension, retention of skills, and maintained a positive attitude toward the use of AR. Pre- and post-intervention comparisons indicated significant differences in weekly progress, long-term retention, satisfaction, willingness to engage, and reduced anxiety levels. Participants also showed marked gains in memory, inferential reasoning, and critical understanding. Overall, the intervention led to significant skill enhancement and long-term maintenance of these educational gains, except for two studies (Tobar-Muñoz et al., 2017; Tosto et al., 2020), which did not report significant differences in performance or average total scores. In these two studies, although performance did not improve significantly, there was an increase in student motivation. These findings suggest that even when measurable performance outcomes are limited, the use of AR can still foster valuable higher-order thinking and engagement.

3.1.1. General Characteristics

Nine studies employed group designs in their interventions (Bursali & Yilmaz, 2019; Cruzado et al., 2020; S. Liu et al., 2024b; Nasongkhla & Sujiva, 2023; Çetinkaya Özdemir & Akyol, 2021; Poobrasert & Luxameevanich, 2023; Şimşek & Direkçi, 2022; Tobar-Muñoz et al., 2017; Tosto et al., 2020) and the remaining two studies employed single-case designs (Lin et al., 2016; Ok et al., 2021).

Single-Case Designs

Each of the two single-case designs lasted approximately three months across all phases. Specifically, Lin and colleagues (2016) employed an ABA design consisting of baseline, intervention, and maintenance phases to investigate the use of mobile AR in enhancing word recognition. The study involved two 11-year-old fifth-grade male students (twins) with ADHD and reading difficulties and was conducted in a resource room. Two teachers and two assistants designed word flashcards and linked them to corresponding videos using the Aurasma app. Participants used a mobile device to scan the flashcards and view the linked videos. Significant reading improvements were observed in both intervention and maintenance phases, supported by effect sizes (Cohen’s d) and confirmed by the Kolmogorov–Smirnov test showing differences between baseline and intervention.

Similarly, Ok et al. (2021) involved three first grade female students, who struggled with reading in a multiple probe across participants design assessing maintenance too. The authors examined video modeling through an AR iPad app to improve phonics skills in three first-grade students with reading difficulties, reporting notable gains maintained for five weeks post-intervention. In addition, two effect sizes (Non-Overlap of All Pairs and Tau-U) were calculated to determine magnitude of the intervention effect.

Group Designs

The study of Tobar-Muñoz et al. (2017) used an AR game developed through a design-based research approach (a research methodology intended to produce design principles and create designs and artifacts for learning) to promote reading comprehension among 51 elementary school students (3rd–6th graders) in a real classroom setting. The researchers developed an AR game for Android tablets and smartphones, inspired by traditional pop-up books, using their illustrations as AR markers. Although the results did not indicate significant difference in reading comprehension compared to traditional methods, the students referred increased interest and motivation during the experience.

The study of Bursali and Yilmaz (2019) involved 89 5th-grade students divided into experimental and control groups to examine the effects of AR on reading comprehension. The experimental group participated in AR-based reading activities, while the control group received traditional instruction. A mixed method, embedded design, was used in this study and the findings indicated that the AR group achieved higher levels of reading comprehension and retention, and reported greater satisfaction, lower anxiety, and positive attitudes toward the use of AR.

Table 1. The reviewed studies.

Reference	Participants/Setting/ Interventionist	Content/ Skill	Type of AR	Objective/ Intervention	Research Design	Results/Findings
Lin et al. (2016)	2M (age: 11 yrs, 5th graders) ADHD and Reading Disabilities Resource classroom/ Teacher	Word reading	mobile AR Aurasma© application (Android and WiFi)	word recognition (60 target words) flash cards & corresponding video: Read the words Write the words Fill the selected word in a sentence 3 months	ABA for single-case research A (3 baseline) B (6 weeks 3 lessons a week 120 min. total) A (3 follow up)	Significant gains obtained for both ‘read the words’ and ‘select the word’ and effects were maintained
Bursali and Yilmaz (2019)	89 (46M, 43F) (age: 10–11 yrs, 5th graders) 43 EG, 46 CG School setting/ Different teachers & researchers	Reading comprehension	Aurasma application in smartphone Animation Pawton software	attitudes towards AR 2 groups & 2 teachers with guidance. 3 weeks: 1 lesson/week	Mixed method (quasi-experimental design & interview)	EG: higher levels of reading comprehension, maintenance and positive attitude towards AR Significant differences Pre-post, per week, permanency, satisfaction, willingness, low anxiety
Tobar-Muñoz et al. (2017)	51 (29M, 22F) (age: 8–12 yrs, 3rd–6th graders) Regular classroom/ Teacher & researchers	Reading comprehension	AR Ole Cierraojos for Android tablets and smartphones	Reading sequentially for 2 weeks in a class with teacher & helpers/researchers An ARGBL experience 2 groups random assignment CG: book only, EG: book & ARGBL	EG and CG	No difference in performance, increase in motivation, Difficulties with technical part Benefits: analyzing or reflecting on the problem. Making hypotheses-proposing/corroborating ideas. Problem solving: arguing, inquiring, debating
Cruzado et al. (2020)	58 (age: 9–10 yrs, 4th graders) 28 EG, 28 CG Pre-established classrooms/ Not specified	Reading comprehension	AR-based mobile application, Mobile-Dagile	Book, AR of the story, glossary, reading, quiz to validate comprehension The influence of AR-based mobile application on reading comprehension levels through stories	EG and CG Pre-post for both groups methodological details on duration & frequency are needed	Significant gains in memory levels, inferential understanding, critical understanding
Çetinkaya Özdemir and Akyol (2021)	54 (EG: 12M, 14F) (4th graders) Not specified/ Not specified	Reading comprehension	Tablets with the mobile AR application	The effect of AR-based reading activities on reading comprehension, reading motivation, attitude towards reading and class participation, and to obtain the students’ views 6 weeks * 5 h per week = 30 h	Mixed method (quasi-experimental design & semi-structured interviews)	Within group differences: high for performance, large for attitude, motivation & medium for participation Between group differences: Large effects for performance, motivation, participation, medium for attitudes
Ok et al. (2021)	3F (age: 6–7.5, 1st graders) Reading difficulties/ at risk for reading disabilities 2 small empty classrooms/ Research team		Video modeling using an AR iPad app	a VM phonics intervention designed with explicit instruction features and administered using an AR iPad app 3 months	single-case multiple probe design across participants design	significant improvement of skills and maintenance of the intervention gains
Tosto et al. (2020)	117 (94M, 23F) (4th, 5th, 6th graders) ADHD (Comorbidities: Dyspraxia, Dyslexia, Language difficulties, Speech difficulties, ASD) At home or in school/ Teachers & parents	Reading, spelling	AHA (ADHD-Augmented) project the Words Worth Learning© (WWL) Programme, a web-based AR learning environment	Supporting the acquisition of English literacy skills	a pre- and posttest design/ 2 intervention groups, WWL-AR vs. WWL programme and one CG	No significant differences in average total scores
Şimşek and Direkçi (2022)	120 students (age: 11–12 yrs) EG: 33M, 27F CG: 30M, 30F Classroom/ Teachers & the researcher	Reading comprehension	The AR storybook application/Android platforms using the Kotlin programming language	6th grade Turkish language coursebooks- 3 texts compare the reading comprehension levels of groups of students who read the same texts in course books through either AR technique or traditional reading technique	a quasi-experimental study with a pre-test-post-test CG Pretest/sensitivity for two weeks +5 weeks trial for each group	Pre-test: No statist. sign. differences Post-test: significant differences at 4 of the 5 variables (except literal comprehension) high comprehension scores/reorganization, inferential comprehension, evaluation and appreciation
Nasongkhla and Sujiva (2023)	32 (3rd graders) Low achievers Classroom/ Teachers or parents	Reading aloud, Reading comprehension	AR-reading platform with 4 components (a PHP database and JavaScript, tools for 2D and 3D media with Unity, marker-based tracking with Vuforia, and worksheets-markers	Reading skills assessment of the AR-reading platform’s efficacy 3 reading levels Individual reading tasks from 9 cat + 1080 from Thai index vocabulary worksheets printed or digital 6 weeks	A research design to analyze the views of learners on an educational product or process	Pre-Post Significant differences for reading aloud and Reading Comprehension

Note. SEN = Special Educational Needs, M = Male, F = Female, yrs = years, min = minutes, EG = Experimental Group, CG = Control Group.

Table 1. The reviewed studies.

Reference	Participants/Setting/ Interventionist	Content/ Skill	Type of AR	Objective/ Intervention	Research Design	Results/Findings
Lin et al. (2016)	2M (age: 11 yrs, 5th graders) ADHD and Reading Disabilities Resource classroom/ Teacher	Word reading	mobile AR Aurasma© application (Android and WiFi)	word recognition (60 target words) flash cards & corresponding video: Read the words Write the words Fill the selected word in a sentence 3 months	ABA for single-case research A (3 baseline) B (6 weeks 3 lessons a week 120 min. total) A (3 follow up)	Significant gains obtained for both ‘read the words’ and ‘select the word’ and effects were maintained
Bursali and Yilmaz (2019)	89 (46M, 43F) (age: 10–11 yrs, 5th graders) 43 EG, 46 CG School setting/ Different teachers & researchers	Reading comprehension	Aurasma application in smartphone Animation Pawton software	attitudes towards AR 2 groups & 2 teachers with guidance. 3 weeks: 1 lesson/week	Mixed method (quasi-experimental design & interview)	EG: higher levels of reading comprehension, maintenance and positive attitude towards AR Significant differences Pre-post, per week, permanency, satisfaction, willingness, low anxiety
Tobar-Muñoz et al. (2017)	51 (29M, 22F) (age: 8–12 yrs, 3rd–6th graders) Regular classroom/ Teacher & researchers	Reading comprehension	AR Ole Cierraojos for Android tablets and smartphones	Reading sequentially for 2 weeks in a class with teacher & helpers/researchers An ARGBL experience 2 groups random assignment CG: book only, EG: book & ARGBL	EG and CG	No difference in performance, increase in motivation, Difficulties with technical part Benefits: analyzing or reflecting on the problem. Making hypotheses-proposing/corroborating ideas. Problem solving: arguing, inquiring, debating
Cruzado et al. (2020)	58 (age: 9–10 yrs, 4th graders) 28 EG, 28 CG Pre-established classrooms/ Not specified	Reading comprehension	AR-based mobile application, Mobile-Dagile	Book, AR of the story, glossary, reading, quiz to validate comprehension The influence of AR-based mobile application on reading comprehension levels through stories	EG and CG Pre-post for both groups methodological details on duration & frequency are needed	Significant gains in memory levels, inferential understanding, critical understanding
Çetinkaya Özdemir and Akyol (2021)	54 (EG: 12M, 14F) (4th graders) Not specified/ Not specified	Reading comprehension	Tablets with the mobile AR application	The effect of AR-based reading activities on reading comprehension, reading motivation, attitude towards reading and class participation, and to obtain the students’ views 6 weeks * 5 h per week = 30 h	Mixed method (quasi-experimental design & semi-structured interviews)	Within group differences: high for performance, large for attitude, motivation & medium for participation Between group differences: Large effects for performance, motivation, participation, medium for attitudes
Ok et al. (2021)	3F (age: 6–7.5, 1st graders) Reading difficulties/ at risk for reading disabilities 2 small empty classrooms/ Research team		Video modeling using an AR iPad app	a VM phonics intervention designed with explicit instruction features and administered using an AR iPad app 3 months	single-case multiple probe design across participants design	significant improvement of skills and maintenance of the intervention gains
Tosto et al. (2020)	117 (94M, 23F) (4th, 5th, 6th graders) ADHD (Comorbidities: Dyspraxia, Dyslexia, Language difficulties, Speech difficulties, ASD) At home or in school/ Teachers & parents	Reading, spelling	AHA (ADHD-Augmented) project the Words Worth Learning© (WWL) Programme, a web-based AR learning environment	Supporting the acquisition of English literacy skills	a pre- and posttest design/ 2 intervention groups, WWL-AR vs. WWL programme and one CG	No significant differences in average total scores
Şimşek and Direkçi (2022)	120 students (age: 11–12 yrs) EG: 33M, 27F CG: 30M, 30F Classroom/ Teachers & the researcher	Reading comprehension	The AR storybook application/Android platforms using the Kotlin programming language	6th grade Turkish language coursebooks- 3 texts compare the reading comprehension levels of groups of students who read the same texts in course books through either AR technique or traditional reading technique	a quasi-experimental study with a pre-test-post-test CG Pretest/sensitivity for two weeks +5 weeks trial for each group	Pre-test: No statist. sign. differences Post-test: significant differences at 4 of the 5 variables (except literal comprehension) high comprehension scores/reorganization, inferential comprehension, evaluation and appreciation
Nasongkhla and Sujiva (2023)	32 (3rd graders) Low achievers Classroom/ Teachers or parents	Reading aloud, Reading comprehension	AR-reading platform with 4 components (a PHP database and JavaScript, tools for 2D and 3D media with Unity, marker-based tracking with Vuforia, and worksheets-markers	Reading skills assessment of the AR-reading platform’s efficacy 3 reading levels Individual reading tasks from 9 cat + 1080 from Thai index vocabulary worksheets printed or digital 6 weeks	A research design to analyze the views of learners on an educational product or process	Pre-Post Significant differences for reading aloud and Reading Comprehension

Note. SEN = Special Educational Needs, M = Male, F = Female, yrs = years, min = minutes, EG = Experimental Group, CG = Control Group.

Cruzado et al. (2020) involved 58 fourth-grade students, aged 9 to 10, from a public school in Lima and used a mobile AR application called ‘IdeAR’ to support reading comprehension through storytelling. A quasi-experimental design with two pre-existing groups was used. The experimental group used the AR-based app, while the control group received traditional instruction. The results indicated that the experimental group showed improvements in memory and content understanding, highlighting AR’s potential as an effective educational tool.

Çetinkaya Özdemir and Akyol (2021) used AR-based reading activities to examine their effects on reading comprehension, reading motivation, attitude toward reading, and class participation for 54 fourth-grade students from a state school in Kars. They used a quasi-experimental design with pretest and posttest measures. A control group was employed. The results indicated significant improvements in the experimental group across all measured variables. The students also reported positive experiences regarding usability, engagement, and skill development.

The study of Tosto et al. (2020) included 117 students with reading difficulties due to a variety of reasons and aimed to examine the effectiveness of an AR-enhanced literacy program for reading and spelling skills. The intervention integrated AR into an existing digital literacy platform. Pre- and post-intervention assessments revealed statistically significant improvements in both reading and spelling performance for the AR group compared to baseline. While the authors did not report explicit effect size metrics, the preliminary findings consistently indicated meaningful gains in literacy outcomes, attention enhancement and task engagement, supporting the benefits of AR integration in literacy instruction.

Şimşek and Direkçi (2022) included 120 students (ages 11–12), divided them into two groups, and assessed their narrative comprehension and retelling skills using pretests and posttests. Over an eight-week intervention period, the experimental group engaged with AR-enhanced book content, whereas the control group received only printed versions of the same texts. Pretest results showed no significant differences between the two groups. However, during the posttest, the AR group demonstrated significantly greater gains in both story comprehension and retelling performance compared with the control group. The experimental group demonstrated significant improvements in higher-order reading comprehension skills, including reorganization, inferential comprehension, evaluation, and appreciation. In contrast, no significant difference was observed in literal comprehension levels between the two groups. Additionally, effect sizes were not reported.

Nasongkhla and Sujiva (2023) report a project and its respective steps (development and implementation) involving 32 students. The researchers developed an AR platform integrating the Picture Word Inductive Model (PWIM) and precision teaching strategies to support reading skill acquisition and to evaluate its effectiveness in improving reading comprehension. The AR system allowed students to use mobile devices to scan textbook pages which activated audio narration, animated visuals and interactive vocabulary prompts. Results from pre- and post-intervention assessments revealed statistically significant improvements in reading comprehension and related literacy outcomes although effect sizes were not reported.

Poobrasert and Luxameevanich (2023) involved 14 students with reading and writing problems. The authors used ‘HadThai’ with AR application to assess its effectiveness in improving students’ spelling and reading abilities, as well as their satisfaction. A pre–post within-group design, measuring students’ literacy skills before and after intervention was employed. Results showed statistically significant improvements in reading and spelling performance at p < 0.05 for all students (first and third graders), indicating notable gains in literacy competency, high levels of satisfaction and preference for the AR-supported learning task. The study, however, does not report specific effect size values, focusing instead on significance thresholds.

S. Liu et al. (2024b) conducted a quasi-experimental study to examine the effects of AR picture books on reading comprehension, story retelling, and reading motivation among 80 primary school students. The participants were randomly assigned to either an experimental group which used AR-enhanced books and scanned story pages to activate multimodal elements, or a control group which read traditional print picture books. Over a three-week period, both groups read the same three Chinese picture books during regular reading classes. Post-intervention results showed that students in the AR group significantly outperformed those in the control group across all assessed outcomes. The authors reported significant differences through t-tests and multivariate analyses of covariance (MANCOVA), although specific effect size values were not provided. Students using AR also reported higher levels of attentional engagement and reading confidence.

3.1.2. Technological Characteristics of AR

All studies specified that they used mobile AR devices (tablets, mobile devices). The authors of nine studies specified the technical development of their AR-based intervention. Two studies (Nasongkhla & Sujiva, 2023; Ok et al., 2021) do not specify the exact AR technologies or software framework used in their study. Nasongkhla & Sujiva, (2023) refer that an AR-reading platform is going to be developed, including multimedia worksheets, and an analytical learning system. Similarly, Ok et al. (2021) only report using video modeling within an AR-based iPad application, without detailing the application’s design or AR features.

The summaries of the studies that specified the technological characteristics of their AR interventions are presented below.

Lin et al. (2016) describe the general procedure regarding the mobile AR application and examine its effects on the learning outcomes of two students with ADHD and reading disabilities. The mobile-based platform employed marker-based AR, utilizing the device’s camera to recognize printed flashcards and overlay multimedia content (3-D animations, visual cues and audio pronunciations). The system provided multisensory feedback (visual, auditory, and tactile) related to animated storytelling, word pronunciation, and interactive prompts to enhance engagement, attention and active learning. The mobile AR environment also offered persistent access, allowing children to revisit target words and animations without repeated marker scans, supporting personalized learning.

Tobar-Muñoz et al. (2017) developed an AR game using a design-based research approach. Physical markers (images or QR codes) triggered digital content (3-D models, animations, visual effects and multimedia elements) when scanned by a mobile device’s camera. The application facilitated interactive gameplay and collaborative learning, while enhancing reading comprehension through immersive storytelling. Thus, the system created a dynamic and meaningful literacy experience that fostered multisensory engagement, representation and action and expression to support inclusive literacy instruction.

Bursali and Yilmaz (2019) implemented a custom marker-based augmented reality (AR) design using the Aurasma platform, integrating rich multi-sensory and interactive features to enhance students’ reading comprehension and long-term retention. Printed reading materials functioned as visual markers that, when scanned with tablets or smartphones, triggered the AR application to display digital content directly over the corresponding text. The smartphone interface was projected onto the interactive whiteboard using the AirDroid application, making the experience visible to all students. The technology was deployed on standard mobile devices, ensuring accessibility and practicality for typical classroom settings without the need for specialized equipment.

In the study of Çetinkaya Özdemir and Akyol (2021), AR technology promoted interactive reading activities through motivation and real-time feedback. The AR system used marker-based tracking within Turkish language materials that, when scanned via mobile devices, triggered immersive multimedia content synchronized audio narration, and interactive visual effects. The audiovisual stimuli and interactive elements helped learners deepen cognitive engagement and supported retention. Persistent accessibility without repeated scanning and gamified storytelling elements accommodated personalized learning.

Cruzado et al. (2020) introduced the mobile application ‘IdeAR’ to improve reading comprehension using the mobile-D-methodology for mobile app development. The authors describe the marker-based AR when scanned via tablet or smartphone cameras, the multimedia content (including 3D animations, visual cues and audio narration) and interactive feedback to enrich storytelling experience. The application’s multisensory feedback, combining the multisensory elements, stimulated deeper cognitive processing and reinforced narrative recall, as evidenced by significant improvements in memory and comprehension among the experimental group. Additionally, its availability on mobile devices ensured flexible and portable access, permitting self-paced exploration beyond the classroom.

Tosto et al. (2020) developed a web-based AR learning environment, namely AHA (ADHD-Augmented), to enhance literacy skills for children with ADHD. The AR system combined marker-based AR triggered by camera-enabled mobile devices to present dynamic, interactive educational content (3-D visuals, animations, audio narration, and real time feedback) directly onto literacy exercises. The AR activities aimed to stimulate attention and motivation, addressing the specific learning needs of students with ADHD.

Şimşek and Direkçi (2022) developed a custom marker-based AR application using the Kotlin programming language for Android tablets. The application enabled students to access AR content (animations, visual and audio effects) seamlessly integrated into the coursebook texts, creating a more immersive and engaging reading experience. The technological features include offline functionality, ensuring accessibility without internet, and seamless integration of AR elements directly into the story texts to enhance reading engagement in classroom.

Poobrasert and Luxameevanich (2023) highlight several characteristics of AR that function as effective assistive tools. The application utilizes marker-based AR, specifically QR codes embedded in physical word cards, which when scanned via mobile devices trigger auditory feedback and visual cues. The use of multimodal output (visual and audio stimuli, and interactive tutorials) supports multiple channels of information processing for students at risk. Additionally, the system’s accessibility and repeatability support consistent vocabulary training, while its child-friendly interface fosters greater student engagement and improves overall usability.

S. Liu et al. (2024b) describe their AR application as a marker-based AR system designed to enhance reading comprehension. The system utilized printed story materials as markers, which, when viewed through a mobile device’s camera, triggered digital content (animations, audio narration and interactive visual effects). The multisensory feedback helped maintain attention, improve comprehension and story retelling accuracy, and reading motivation. Delivered via mobile devices, the AR books provide flexible access and self-paced exploration, enabling personalized learning. The integration of multimedia elements, such as 3D animations and audio narration, further enriched the storytelling experience making it more immersive and engaging for young readers.

3.1.3. Universal Design for Learning

Across the reviewed studies, the technological characteristics of AR, including marker-based systems, mobile device compatibility, and multimodal feedback, were intentionally designed to create engaging, interactive, and accessible literacy learning environments. Through animated storytelling, auditory word pronunciation, interactive prompts, or collaborative gameplay, these AR applications supported key elements of reading development such as comprehension, attention, motivation, and active participation. By leveraging AR’s inherent features (e.g., contextual augmentation, multisensory stimulation, interactivity, and personalization) the systems enabled flexible, meaningful and student-centered learning experiences that accommodated a range of educational needs, particularly among learners with ADHD, reading disabilities, or students at risk.

The applications’ user-friendly interfaces further supported independent exploration, allowing students to learn at their own pace. This multisensory, self-paced design aligned with principles of accessibility and engagement, central to supporting learners with reading and writing difficulties. In line with research question 2, the UDL framework (CAST, 2018) provided a useful lens for evaluating how these interventions addressed learner variability. UDL, originally developed by CAST in the 1990s and grounded in cognitive neuroscience, emphasizes the proactive design of flexible learning environments that minimize barriers and maximize accessibility for all learners. The framework is structured around three core guidelines: multiple means of engagement (the “why” of learning), multiple means of representation (the “what” of learning), and multiple means of action and expression (the “how” of learning).

Aligned with these principles, the AR interventions in the reviewed studies provided: (i) multiple means of engagement through game-like features, interactive tasks, and AR prompts that sustained motivation and encouraged persistence, particularly for learners who might otherwise disengage from print-based tasks. (ii) multiple means of representation by presenting text alongside animated visuals, synchronized audio, and multimodal scaffolds that supported comprehension beyond decoding alone, (iii) multiple means of action and expression by enabling students to respond verbally, interact with AR markers, retell narratives, and engage in collaborative dialogue, offering diverse avenues for demonstrating understanding. Collectively, these findings underscore AR’s potential as a powerful assistive and inclusive technology for literacy instruction. By embedding UDL principles into their design, AR applications not only enhanced motivation and comprehension but also supported equitable participation for diverse learners, including those with reading difficulties.

The app’s friendly interface supported independent exploration, allowing students to work at their own pace. This multisensory, self-paced procedure aligned with principles of accessibility and engagement, key to supporting learners with reading and writing difficulties. Aligned with the principles of UDL, these AR interventions provided multiple means of engagement (e.g., game-like features and interactive tasks or AR prompts and animations), multiple means of representation (e.g., animated visuals, synchronized audio, and printed text), and multiple means of action and expression (e.g., verbal responses, scanning interactions, narrative retelling and collaborative dialogue). Collectively, these studies underscore the potential of AR as a powerful assistive and inclusive technology for literacy instruction, capable of fostering motivation, enhancing comprehension, and supporting equitable participation for diverse learners.

3.2. What Are the Advantages and Disadvantages of These Studies Based on Quality Indicators?

In 2003, the Division for Research of the Council for Exceptional Children outlined four primary research design types used in special education: experimental group designs, correlational group designs, single-subject designs, and qualitative designs. To assess the effectiveness of practices within each type, they proposed a set of quality indicators (Gersten et al., 2005). Building on this framework, Cook and Cook (2016) offered a comparable classification system, grouping research designs into four categories: descriptive, relational, experimental, and qualitative.

The quality of single-subject research is assessed using seven key indicators, which address elements such as participant selection, variable definition, baseline conditions, internal validity, and social validity (Horner et al., 2005). Similarly, group experimental and quasi-experimental studies are evaluated using a set of essential and recommended indicators focusing on participant characteristics, intervention procedures, outcome measures, and data analysis methods (Gersten et al., 2005).

Two studies employed single-subject design (Lin et al., 2016; Ok et al., 2021). Both met most of the seven quality indicators proposed by Horner et al. (2005), demonstrating strong methodological rigor. The rest nine studies implemented a group experimental or a quasi-experimental research design (Table 1). Five of them included a control group (Cruzado et al., 2020; S. Liu et al., 2024b; Şimşek & Direkçi, 2022; Tobar-Muñoz et al., 2017; Tosto et al., 2020) together with pre and post-tests to measure learning outcomes.

To provide an evaluation of these studies the criteria presented by Gersten et al. (2005) have been followed. The Essential Quality Indicators they proposed include adequate information about sampling procedures, sufficient details about the implementation of intervention and outcome measures, as well as good presentation of the results and data analysis. The Desirable indicators involve eight criteria which assess 1. attrition details, 2. reliabilities of measures and procedures including scorers’ familiarity with intervention study conditions, 3. a follow-up measure, 4. validity of measures (criterion-related, construct), 5. fidelity indicators (e.g., procedures specified, time allocated etc), 6. instruction of comparison group, 7. evidence (e.g., audio, video, photos) capturing the nature of intervention, 8. clear presentation of study. The authors propose a scheme for ranking (high quality or acceptable). To be considered high quality the study must meet all but one of the Essential Quality Indicators and at least four of the quality indicators in the Desirable range. A study of acceptable quality must meet all but one of the Essential Quality Indicators and at least one of the Desirable range.

All of them provided basic information as regards sampling procedures and description of experimental subjects and control group when present. The studies were selected on the condition that there was sufficient information on the progress of the students and measures to indicate that (e.g., achievement pre-post intervention). Presentation of effect sizes was not a criterion for selection. Likewise, most of the studies reported basic information on the researchers’ efforts to ensure the quality of the outcome measures. All studies could benefit from the addition of more information as regards implementation of intervention.

Table 2. Quality indicators based on Gersten et al. (2005).

	Essential Quality Indicators				Desirable	Ranking
	Participants Sampling	Implementation of Intervention	Outcome Measures	Quality Indicators for Data Analysis	Criteria Max 8	Acceptable or High Quality
1. Lin et al. (2016) *	+	+	+	+	2, 3, 4, 8	High
2. Tobar-Muñoz et al. (2017) †	+	+	+	+	2, 6, 7, 8	High
3. Bursali and Yilmaz (2019) †	+	+	+	+	2, 3, 7, 8	High
4. Cruzado et al. (2020) †	+			+	7, 8, 1	-
5. Çetinkaya Özdemir and Akyol (2021) †	+		+	+	6, 8	Acceptable
6. Ok et al. (2021) *	+	+	+	+	3, 4, 5, 8	High
7. Tosto et al. (2020) †	+		+	+	8	Acceptable
8. Şimşek and Direkçi (2022) †	+	+	+	+	4, 7, 8	High
9. Nasongkhla and Sujiva (2023) †	+		+	+	1, 8, 4	Acceptable
10. Poobrasert and Luxameevanich (2023) †	+		+	+	2, 8	Acceptable
11. S. Liu et al. (2024b) †	+	+	+	+	4, 6, 8	Acceptable

Note. * studies with single-case research design, criteria also considered Horner et al. (2005). † group design studies; criteria from Gersten et al. (2005).

shows that about half of them provided a good account of steps and procedures. Based on the Essential Quality Indicators only one study, Cruzado et al. (2020), failed to meet the requirements for both the implementation of the intervention and the description of the outcome measures. Therefore, it could not be characterized as acceptable.

Table 2 shows how the eleven studies meet desirable indicators. All of them had been written in a reasonably clear manner (8), whereas only some (n = 4) provided sufficient material e.g., photos to accompany intervention materials and methodological details, aiding understanding of the procedures followed (7). Information on the instructions provided to the comparison group was reported in three studies (n = 3) and mentioned in broad terms (6). Only one study Ok et al. (2021) provided information regarding treatment fidelity (5), and criterion-related or construct validity was allocated when there was clear indication that materials had been taken from school books or experts had contributed to their development (n = 4). Two studies reported a follow up (3) and two provided an account of attrition rates (1). Five studies were allocated criterion (2), when there was an indication that professional teachers carried out the intervention and the researchers did not interfere with procedures due to their familiarity with study conditions. Since research quality has broader educational implications, Gersten et al. (2005) stated that one criterion for identifying a practice as ‘promising’ is its support by at least four studies of acceptable quality or two high-quality studies. However, only in the presence of adequate effect sizes can the condition be fully met. As shown in Table 2, there are five studies designated as acceptable and five designated as high quality, which may constitute enough evidence to introduce a meaningful involvement of AR in the primary school reading curriculum.

In summary, the body of studies reviewed reflects several notable strengths, including adherence to established quality indicators, adequate reporting of participant characteristics and outcome measures, and, in many cases, the inclusion of control groups with pre- and post-tests, which enhances internal validity. These features provide a degree of methodological rigor that supports cautious confidence in the reported outcomes. At the same time, the studies also present weaknesses that temper the strength of the evidence base. Limited attention to intervention fidelity, and follow-up measures, alongside inconsistent reporting on comparison group instruction restricts the reliability and generalizability of the findings. Moreover, the lack of consistent effect size reporting hinders an assessment of the practical significance of the interventions. Taken together, these strengths and limitations highlight that while the current evidence is promising, further high-quality research with more comprehensive methodological reporting is required to establish a stronger and more conclusive knowledge base.

3.3. What Are the Limitations of These Studies That May Introduce Threats to Their Validity?

There are threats to the validity of any study employing quasi-experimental design. Intervention studies cannot easily avoid pitfalls that limit the validity of their outcomes due to constraints in their implementation. Problems with time, space, personnel, administration, and coordination constitute some of the most frequent sources of difficulty when targeting an intervention study.

Ideally, experimental and control groups should undergo matching on certain critical variables and control of their reading level. The only study which statistically tested equality on reading level was the one by Bursali and Yilmaz (2019). However, to ensure feasibility of the study, whole classes have been employed, a method commonly used across the reviewed studies. In addition, most of them did not include a control group or subjects. Participation of controls (subjects or groups) in the study cannot be warranted since their availability is uncertain and may be influenced by external factors beyond the researchers’ control.

Moreover, it cannot be precluded that the outcomes may reflect a teacher’s effect, the ability of different tutors to accomplish the task of reading instruction, which may also be a source of variation across the conditions and settings. Without proper monitoring and standardized training, it may be difficult to discriminate to what extend individual teacher or instructor qualities may be responsible for the results. Only the individual attention that is provided particularly in the case study designs would be enough to generate an effect that could not be attributed solely to experimental conditions. Thus, fidelity of implementation remains a critical concern.

Technology-based interventions require mastery of the equipment employed and it has the potential to affect outcomes since not all children can cope with technology equally well. To address this, Tobar-Muñoz et al. (2017) engaged the teacher and three researchers present in the class to help with technical questions and problems. This, however, deviates from the regular classroom ecology.

Novelty effects may also pose a threat to internal validity, as participants’ initial enthusiasm increased motivation to engage with the new or unfamiliar intervention. This can temporarily inflate performance outcomes. This effect, however, may diminish over time as the novelty wears off, making it difficult to distinguish between genuine learning gains and temporary motivational sparkles. Therefore, it is essential to interpret short-term improvements with caution. To evaluate the sustainability of observed effects follow-up assessment is therefore important to find in an intervention technology-based study.

Assessment procedures should also be carefully monitored providing sufficient information on how these were developed and appropriately implemented to minimize sensitivity to experimental conditions. Furthermore, generalizability of the positive effects to standardized reading tests or non-experimental reading materials remain uncertain, as the evidence emerging from these studies is negligible.

Overall, the limitations of the included studies reveal several common threats to validity. The use of quasi-experimental designs without adequate matching or control groups limits internal validity and raises concerns about selection bias. Variation in teacher effects, fidelity of implementation, and technological challenges introduce additional sources of confounding that may obscure the true impact of the interventions. Short-term novelty effects further complicate interpretation, particularly when follow-up assessments are absent. Finally, weaknesses in the design and reporting of assessment procedures, combined with the limited evidence of transfer to standardized measures, restrict both the reliability and generalizability of the findings. These threats underscore the need for more rigorous study designs, systematic fidelity monitoring, and long-term evaluation to strengthen confidence in the outcomes of technology-based reading interventions.

4. Discussion

This review aimed to evaluate the effectiveness of AR-based interventions on reading skills of children in primary education up to age 12. The study focused on first language reading processes and children with and without reading difficulties. The findings suggest that the emerging field of AR interventions holds significant potential for enhancing the reading skills of students in primary education.

The first research question aimed to assess the effectiveness of AR-based interventions in improving reading skills. Increased learning outcomes were observed in academic tasks, particularly in reading and reading comprehension. Given the well-established link between cognitive functions, such as attention and memory, and academic achievement, AR-based interventions may enhance not only school performance but also support greater school inclusion and improve children’s peer relationships.

It is important to understand what motivates and engages primary school students to participate in AR learning experiences. AR learning environments can be co-developed with students or teachers who reflect the needs of the intended users, which may enhance their relevance, accessibility, and educational impact. In AR applications for primary education, similar user-centered design approaches are emerging, allowing students to interact with digital content overlaid in real-world settings. Although preliminary findings across several studies suggest that AR can support improved attention, knowledge retention, and enjoyment, much of this research has relied on small-scale or non-experimental designs, limiting the generalization of the outcomes.

Furthermore, more rigorous methods are needed to evaluate the effectiveness of AR tools, including assessments of learning transfer and cognitive load, as well as the psychometric properties of the evaluation instruments. Based on well designed and executed studies teachers and researchers can make informed decisions about the use of AR in literacy instruction. An evaluation such the one addressed in the present review can help teachers and researchers to identify paradigms worth following and implement their core elements for a meaningful in-class intervention. Validity issues weaken the credibility of findings and poor validity may lead to wasted resources, false expectations and frustration. From a researcher’s perspective, AR shows considerable promise, though more evidence is needed before it can be fully endorsed as an evidence-based practice. Overstating AR’s benefits could mislead policy decisions, curriculum changes and funding priorities.

In addition, expert input on the pedagogical structuring of AR environments is crucial. For instance, systems based on AR illustrate how educational settings can be aligned with specific curriculum goals. AR can offer similarly rich contextual learning opportunities if grounded in both learner input and sound instructional design principles.

In line with Mayer’s cognitive theory for multimedia learning, AR users often experience reduced cognitive load, increased motivation, and more positive attitudes toward learning when engaging with AR-enhanced texts. Such benefits contribute to a more stimulating and effective learning environment since information is conveyed through multiple sensory channels and instruction is better tailored to individual abilities and preferences. This makes AR particularly effective in addressing diverse learning needs and advancing educational inclusivity. AR has shown promise for learners who face challenges with traditional, text-based instruction. Embedding vocabulary in authentic context-rich scenarios and situating language within meaningful, real-world contexts, AR provides visual and situational cues that support both understanding and long-term retention of linguistic information (Weerasinghe et al., 2022). This approach resonates with Vygotsky’s Sociocultural Theory, which emphasizes the role of social interaction, context, and cultural tools in shaping cognitive development (Ferreira et al., 2021). From this perspective, learning is a socially embedded process, enriched by experiences that connect new knowledge with prior understanding, allowing students to integrate new vocabulary into their own cognitive frameworks and lived experiences. This not only deepens comprehension and word recognition but also enhances the transfer of knowledge across different contexts and disciplines.

These findings also resonate strongly with the principles of the Universal Design for Learning (UDL) framework (CAST, 2018), which emphasizes the importance of offering multiple means of engagement, representation, and action/expression. AR’s multimodal nature appears particularly well suited to supporting UDL principles in literacy contexts. For instance, interactive features and gamified elements align with multiple means of engagement, sustaining motivation for students who may otherwise disengage from print-based tasks. The integration of synchronized audio, animated visuals, and textual overlays provides multiple means of representation, ensuring that content is accessible to students with different perceptual or decoding challenges. Finally, AR tasks such as scanning, retelling, or collaborative dialogue create multiple means of action and expression, giving learners flexible opportunities to demonstrate comprehension beyond written responses.

By mapping the reviewed interventions against UDL guidelines, this study highlights how AR can serve not only as a technological innovation but also as an inclusive pedagogical tool. Embedding AR within a UDL-informed instructional design offers a pathway for addressing learner variability in primary education, especially for students with ADHD, reading difficulties, or other learning needs. Future research should therefore examine not only the effectiveness of AR in improving reading outcomes, but also the extent to which its design explicitly incorporates UDL principles to enhance accessibility, reduce barriers, and promote equitable participation.

4.1. Disadvantages

Despite its promising advantages, the integration of AR in education presents several challenges. Some studies raise concerns about its potential impact on traditional reading habits. Parents may react to the number of hours children spend on screen. Moreover, the complexity of AR environments can pose significant learning barriers, particularly for younger students. Dunleavy et al. (2009) found that learners participating in multi-user AR simulations frequently reported feelings of confusion and disorientation. These challenges often stem from the cognitive strain involved in navigating unfamiliar technologies while simultaneously engaging with demanding academic content. In the absence of sufficient instructional scaffolding, the dual task of mastering both interface mechanics and learning objectives can detract from AR’s educational effectiveness.

The effective adoption of AR in primary education depends not only on technological quality but also on learners’ developmental readiness, thoughtful pedagogical design, and the availability of structured guidance and training. By carefully addressing these factors, educators can better leverage AR’s transformative potential while moderating its limitations.

Despite these strengths, many AR systems reviewed lacked a clearly articulated theoretical framework linking technological features to learning outcomes. For instance, while many interventions incorporated multimedia elements (text, audio, animation), they rarely referenced principles from established learning theories, such as Mayer’s Cognitive Theory of Multimedia Learning, to justify design choices. Similarly, few studies systematically examined whether AR’s interactive features support transfer of learning across contexts which is critical for reading development.

Additionally, there is a lack of standardization in AR system design across studies. Some applications offer simple overlays of vocabulary items; others simulate entire narrative scenes. This variation presents a challenge for drawing generalizable conclusions about which design characteristics most effectively support early literacy skills. As shown in Table 1, the AR platforms range widely in interface type (e.g., mobile apps, tablet-based books, web-based AR), content complexity, and degree of learner control. This fragmentation also makes it difficult to assess whether AR interactions are intentionally designed to optimize cognitive processing or simply intended to boost engagement. The potential for cognitive overload remains a concern, especially in systems where visual and auditory stimuli are abundant but poorly integrated (Dunleavy et al., 2009).

Finally, while some systems were developed with user-centered or participatory design approaches, most were created by researchers or developers without direct input from the student users or classroom teachers. This gap may reduce the ecological validity and classroom applicability of these tools.

4.2. Pre-Intervention Training

Prior to the intervention phase, several studies conducted a structured training session to familiarize participants with the mobile AR system. This involved demonstrating the targeted reading process (e.g., from scanning the printed character cards and triggering the associated animation and audio feedback, to responding to on-screen prompts). Students then completed the practice trials under supervision, during which the interventionists provided immediate guidance on correct handling of the device and interpretation of feedback. Clear instructions were given on how to respond to errors, when to repeat trials, and how to request help, thereby ensuring consistent participation and reducing task-related confusion.

The pre-training phase offers several important advantages, particularly in studies involving students with ADHD and reading disabilities. By explaining and demonstrating the intervention process beforehand, the teacher reduces anxiety, increases student confidence, and helps establish clear expectations for how to interact with the mobile AR system. This initial guidance ensures that participants understand how to use the technology correctly, which minimizes distractions and technical confusion during the actual learning tasks. For students with attention and learning challenges, structured preparation can improve focus, enhance engagement, and support better task persistence. Moreover, it helps standardize the participants’ starting point, reducing variability due to unfamiliarity with the tools or procedures, an important factor in maintaining the validity and reliability of the research outcomes. Overall, the pre-training phase creates a smoother transition into the intervention, supporting both usability and learning effectiveness.

5. Implications for Practice

The findings of this review offer several practical implications for educators, instructional designers, and researchers who implement AR in literacy instruction for primary education students, particularly those with reading difficulties or at risk of academic underperformance.

First, AR can serve as an effective inclusive technology, especially when designed considering learner’s specific profile and educational needs. Features such as multimodal cues (visual, auditory, and interactive feedback), adaptive pacing, and intuitive interfaces support individual learning and foster independence. Educators should therefore consider AR tools that incorporate Universal Design for Learning guidelines, offering multiple means of engagement, representation, and expression to accommodate a wide range of literacy needs.

Second, AR applications allow students to be engaged in educational activities where they can repeatedly practice their skills in a safe and controllable environment that integrate storytelling, gamified tasks, and sensory-rich environments can significantly enhance student motivation, attention, and comprehension. These tools are particularly beneficial for learners with ADHD, dyslexia, or other reading challenges, as they create immersive and stimulating contexts for reading practice.

Third, the design of AR interventions should be pedagogically grounded. While the reviewed studies demonstrated positive outcomes, many lacked a clearly theoretical foundation. AR tools that are guided by constructivist or experiential, inquiry-based, or social-emotional learning theories may lead to deeper engagement. To maximize impact, educators and developers should align AR content with more tailored and meaningful literacy experiences. Integrating frameworks such as Mayer’s multimedia learning principles and Vygotsky’s sociocultural theory can help ensure that AR features support cognitive and social development.

6. Limitations and Future Research

Our review revealed several key limitations in existing studies. Only a limited number of studies on AR-based reading interventions for primary education students with reading difficulties or those at risk were identified in this review. A key limitation was the restriction to studies published in English and in peer-reviewed journals or conference proceedings, which may have excluded relevant research published in other languages or through alternative academic sources. Expanding future searches to include additional databases, grey literature, and non-English sources could enhance the comprehensiveness of the evidence base. Another concern is the limited generalizability of findings, often stemming from small sample sizes and context-specific implementations. Many studies also lack sufficient experimental rigor, undermining the reliability of their conclusions. Furthermore, most interventions are short-term with minimal follow-up, making it difficult to assess long-term impact. Future research should prioritize the integration of established learning theories and pedagogical frameworks to guide the design, implementation, and evaluation of AR interventions, ensuring they are both educationally grounded and practically effective.

7. Conclusions

This review highlights the growing potential of AR in supporting reading development among primary students, particularly those with reading difficulties or at risk. AR’s unique features, such as interactivity, multimodal feedback, and personalization, can enhance motivation, comprehension, and engagement in literacy learning. However, while the results are promising, many interventions lack a strong pedagogical foundation. To maximize educational value, future AR applications should be grounded in established learning theories and aligned with instructional goals. When thoughtfully integrated, AR can serve as a powerful, inclusive tool to enrich literacy instruction and support diverse learners.