Research foci, design features, and accessibility challenges in digital mathematics and geography textbooks: a scoping review

Digital textbooks are becoming increasingly important in mathematics and geography lessons, as interactive and multimedia features that go beyond the capabilities of printed resources are believed to offer particular learning benefits. Despite this growing significance, there has so far been no systematic research review that examines digital textbooks as a subject in their own right in these disciplines. The issue of accessibility, in particular, has been largely overlooked in previous textbook research and has gained particular relevance due to current legislation. This scoping review therefore aims to map the existing literature and to answer three questions: what are the main research foci, which design features of digital textbooks are being investigated, and in what way and to what extent is accessibility addressed. Based on a systematic search in Web of Science and Scopus following the PRISMA-ScR framework, 78 studies (73 in mathematics, five in geography) were included in the final analysis corpus and evaluated using a common coding guide. The results reveal a young, rapidly growing, yet unevenly distributed field of research. Studies in the field of mathematics substantially outnumber those in geography, and the research focus lies predominantly on the development, production and use of digital textbooks, whilst effectiveness and content analyses are comparatively rare. At the feature level, research concentrates primarily on immediately visible functions such as text or visualizations, whilst system-based and adaptive functions are scarcely investigated. Although accessibility is clearly present in the literature, it is conceptualized inconsistently. The review highlights key research gaps, in particular the disciplinary asymmetry between subjects and the need for an explicit examination of accessibility, thereby providing a basis for future research and the inclusive design of digital textbooks.

Digital textbooks in mathematics and geography education

In mathematics and geography lessons, as in many other school subjects, textbooks are considered one of the central teaching and learning resources (Bagoly-Simó, 2021; Fan et al., 2013; Rezat et al., 2021). They fulfil a mediating function between the intended and implemented curriculum and thus influence both the what and the how of teaching, i.e., what content is covered and how this is done (Cai and Howson, 2013; Chazan and Yerushalmy, 2014; Valverde et al., 2002). At the same time, the increasing digitalization of society and the education system is also changing the conditions under which textbooks are designed, produced and used (Behnke, 2021; Choppin and Borys, 2017). Consequently, digital textbooks are becoming increasingly important. As so-called second-generation digital textbooks, i.e., those that are not merely digitized versions of printed textbooks, they are characterized by distinctive digital enhancements that go beyond what printed materials have to offer (Pepin et al., 2015). Following Brnic et al. (2024), digital textbooks are understood as technology-based curricular resources for learning and teaching (mathematics and geography), which are distinguished from printed textbooks by enhancement features such as interactive and multimedia elements, feedback options, and adaptivity and customization functions. Such enhancement features are regarded as key potential benefits of digital textbooks in both mathematics and geography teaching contexts (see Behnke, 2021; Choppin et al., 2014; Rezat, 2025). Both subjects depend to an unusual degree on non-textual representations, mathematics on symbolic notation and graphs, geography on maps and layered spatial information. Printed pages can support such content only partially, which makes digital enhancement features especially consequential in these two subjects.

These enhancement features are considered to offer considerable opportunities for learning. In particular, interactive elements, dynamic visualizations and computer-based feedback are considered promising features of digital textbooks that can positively influence learning (Berney and Bétrancourt, 2016; Castro-Alonso et al., 2019; Moreno and Mayer, 2007; van der Kleij et al., 2015). Meta-analyses from instructional psychology suggest that learners benefit more from dynamic than static visualizations (Berney and Bétrancourt, 2016; Castro-Alonso et al., 2019), a finding that has also been demonstrated for specific areas of mathematics (e.g., Nguyen and Greefrath, 2025). Additionally, computer-based feedback is considered one of the most effective forms of feedback (van der Kleij et al., 2015). Interactivity is considered particularly conducive to learning, as it enables learners to actively engage with learning content, whether by retrieving feedback on their own solutions (dialoguing), controlling their own learning pace (controlling) or directly manipulating representations (manipulating) (Moreno and Mayer, 2007). Since deep learning depends on cognitive activity, interactive learning environments support learners in actively constructing knowledge rather than passively receiving information (Moreno and Mayer, 2007). Hillmayr et al. (2020) were able to show in a meta-analysis that the use of interactive digital tools in secondary school mathematics and science lessons has a positive overall effect on student performance (d = 0.65). Direct comparisons between digital and printed textbooks also show positive effects on learning achievement (e.g., Brnic et al., 2024; Kaplar et al. 2022; Reinhold et al., 2020).

However, whether and to what extent these promising features are actually implemented in specific digital textbooks is a separate question. Textbook analyses show that the potential described above is often not yet being exploited in current digital textbooks. Interactive elements, opportunities for collaboration and customization options are rarely found in digital mathematics and geography textbooks, and the task formats vary only slightly from their printed counterparts (e.g., Behnke, 2021; Behnke and Bagoly-Simó, 2016; Brnic and Greefrath, 2025; Glasnovic Gracin and Kristo, 2022).

Nevertheless, the existence of digital enhancement features alone does not allow any conclusions to be drawn about the actual quality of a specific digital textbook. This is evident, for example, in the fact that, despite a growing range of digital textbooks in Germany, clear, subject-specific quality criteria have hardly been developed and established as binding standards (Rezat, 2025). Elaborate educational concepts that consistently exploit the possibilities of digital technologies for learning and teaching geography also remain a desideratum (Behnke, 2021; Behnke and Bagoly-Simó, 2016). Furthermore, the actual use of digital textbook elements is a necessary precondition for didactic potential to become effective at all (Pepin and Gueudet, 2018).

In order to systematically identify which aspects of digital textbooks have been the focus of research to date and where gaps exist, a structured research overview is needed, but this is currently lacking in the field of digital textbooks. Despite its growing relevance, research on digital textbooks has not been systematically explored in either mathematics education or geography education through a separate review.

Textbook research in geography has a long tradition. In an overview of studies published in six international and German-language journals from 1960 to 2020, Bagoly-Simó (2021) shows that the literature has been dominated for decades by two main issues: the development and validation of evaluation criteria for textbooks on the one hand, and the analysis of textbook content on socially relevant topics on the other. Studies that refer to the design, production and use of textbooks from a comparative perspective, on the other hand, are substantially underrepresented. Digital textbooks are not discussed in this overview study (Bagoly-Simó, 2021).

In mathematics education, there are also recurring overview studies, but these either address textbook research in general, i.e., covering both printed and digital textbooks, or treat digital textbooks merely as a subcategory of a broader field of curriculum resources (see Fan et al., 2013, 2025; Rezat 2024; Rezat et al., 2021). Fan et al. (2013) characterized the field of mathematics textbook research as a dynamically developing but still unbalanced field of research, in which around one third of studies focused on textbook analyses, followed by comparative studies and usage studies. Research on digital textbooks did not yet form a separate category at that time, although the authors already anticipated its growing importance. In recent years, further overview studies have been added in the field of mathematical education, which help to classify the field. There has been an increasing shift towards a broader understanding of textbooks as part of a larger spectrum of curriculum resources that are used as materials in the classroom (Leung et al., 2023). In his survey of research on curriculum resources for the period 2018 to 2023, Rezat (2024) found that digital mathematics textbooks account for only about 16% of the studies included. Similar to Fan et al. (2013), content-related textbook analyses dominate the field, accounting for more than half of the textbook studies. Just under a quarter of the studies deal with the use of textbooks, around 14% examine effects on students or teachers, and around 11% are explicitly devoted to the design of textbooks. No systematic differentiation is made between digital and printed textbooks. In a micro-scoping review, Fan et al. (2025) focus on textbooks in relation to curriculum development and teaching reform, explicitly pointing to the need for further research on the quality assessment of digital learning resources.

To date, there has been no review that focuses on digital textbooks as an independent subject of investigation in mathematics and geography education. In addition to questions about content focus and didactic design features, the question also arises as to the extent to which digital textbooks are equally accessible to all learners. This is an aspect that has been largely ignored in previous textbook research and to which this review therefore devotes particular attention.

Why accessibility matters: from technical standards to inclusive design

Accessibility has increasingly been recognized as a central quality criterion for digital educational materials, rather than a peripheral or optional concern. In the context of digital textbooks, accessibility refers to the extent to which learners with diverse abilities can use and benefit from the material without barriers. Such barriers may be sensory (e.g., relating to vision or hearing), motor, cognitive, or linguistic in nature. Importantly, accessibility is not only a matter of compliance with technical standards. It also concerns whether the design of a textbook supports meaningful participation in learning and reduces barriers to understanding and engagement, an understanding rooted in inclusive education and Universal Design for Learning (UDL) (CAST, 2024).

This shift is reinforced by a clear policy context in Europe. The European Accessibility Act (European Parliament and Council of the European Union, 2019), whose accessibility requirements became applicable in 2025, establishes harmonized requirements for a range of digital products and services, including e-books, thereby strengthening the expectation that digital learning materials should be accessible from the outset. Complementing this regulatory framework, widely used guidelines such as the Web Content Accessibility Guidelines (WCAG) articulate four core principles, content should be perceivable, operable, understandable, and robust, and provide a shared reference point for assessing technical accessibility (see World Wide Web Consortium, 2024). For digital textbooks, adherence to these principles may involve, for example, compatibility with assistive technologies such as screen readers and text-to-speech, alternative text for images, keyboard navigation, and structured markup that supports orientation and comprehension.

For the purposes of this review, accessibility is treated as a multi-dimensional construct that encompasses both technical accessibility and pedagogical accessibility (e.g., CAST, 2024; Silva et al., 2025). Technical accessibility concerns the compatibility of materials with assistive technologies and their conformity with formal accessibility requirements. Pedagogical accessibility, by contrast, concerns whether design choices reduce barriers to understanding and engagement, for instance, by offering flexible representations, multiple pathways for engagement, and varied options for expression, in line with UDL. This framing allows the review to capture not only whether accessibility is addressed, but how it is understood and integrated into research on digital textbooks.

Mathematics and geography as critical cases for accessibility

Mathematics and geography were selected as critical cases because both disciplines rely heavily on non-textual, spatial, and symbolic representations that are difficult to linearize and therefore particularly prone to accessibility barriers in digital formats (see Brnic et al., 2025). Unlike running text, content such as equations or maps cannot simply be read out sequentially. It requires deliberate structuring to remain perceivable, operable, and understandable for learners with diverse needs. This shared characteristic makes the two subjects especially informative for examining how accessibility is addressed in research on digital textbooks.

In mathematics, textbooks typically include complex notations, graphs, and interactive problem-solving environments. Here, accessible markup such as MathML is essential for screen reader compatibility (World Wide Web Consortium, 2024), while interactive elements raise additional requirements regarding keyboard operability and the handling of dynamic content. In geography, textbooks commonly incorporate maps, spatial visualizations, and layered geographic information, which pose specific accessibility challenges due to their inherently visual and spatial nature. Research on digital map accessibility shows that such representations often lack adequate text alternatives, clear navigation structures, and compatibility with assistive technologies, thereby limiting access for learners with visual impairments (e.g., Biggs et al., 2025; Hasan and Gjøsæter, 2021). Beyond sensory access, both subjects also place demands on cognitive accessibility, as their high visual information density and the simultaneous processing of multiple representational layers can themselves constitute barriers to understanding.

These characteristics are not unique to mathematics and geography. Other subjects face comparable demands and would be equally suitable candidates. The present selection follows the scope of the Digital Education Material (DEM) project, an Erasmus + initiative developing guidelines for accessible digital textbooks in line with UDL principles and WCAG standards (Brnic et al., 2025), in which mathematics and geography are the participating subjects. They were included there for the reasons outlined above. Their representational demands make accessibility requirements particularly tangible, while recurring in comparable form in other subjects. The two subjects are therefore treated as illustrative rather than exhaustive cases. By focusing on mathematics and geography, the review maps how accessibility has been operationalized in subject-specific research and identifies barriers and practices relevant for improving digital textbook design in line with European accessibility goals.

This paper is designed as a scoping review. Scoping reviews are a form of knowledge synthesis that aims to systematically map the scope, range and nature of the available evidence on a topic and to identify key concepts, theories and research gaps (Horsley, 2019; Tricco et al., 2018). The field of research is characterized by a high degree of heterogeneity in terms of methods, disciplines and research questions, which corresponds to a constellation for which scoping reviews are expressly recommended (Tricco et al., 2018). The review was conducted and reported in accordance with the guidelines of PRISMA-ScR (Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews) (Tricco et al., 2018).

Search strategy and selection criteria

The literature search was conducted in Web of Science and Scopus, which were selected because their combined coverage spans the relevant disciplines and minimizes the risk of overlooking pertinent studies. Both searches implemented the same agreed search logic, combining a resource block covering the terms used for digital textbooks with a subject block covering the two subject areas: (((digital AND textbook) OR e-textbook OR electronic textbook) AND (math OR geography)). The two blocks were combined with the Boolean operator AND and the terms within each block with OR, so that mathematics and geography were treated as alternative subject terms rather than as a required combination. No database-side filters or limits were applied with regard to document type, publication year, language or subject area. All eligibility criteria were instead applied during screening. The complete search strings for both databases, together with the fields and editions searched, are documented in Supplementary File S1. In addition to the institutional search at the University of Münster, a cross-validation search in Web of Science was carried out at the Free University of Bozen-Bolzano to ensure the most complete coverage of the relevant literature. Because index coverage differed between the two institutional subscriptions, this search contributed 61 records that were not contained in the export from the University of Münster, so that 362 unique Web of Science records entered the review alongside 575 Scopus records. The records exported from both databases and from both institutional searches were compiled in a single spreadsheet. Duplicates were then identified and removed manually. To include publications up to and including 2025, the final literature search was conducted on 25 February 2026.

The inclusion criteria were: (1) publications up to and including 2025, (2) articles from scientific journals and conference proceedings, (3) studies that explicitly deal with digital textbooks in school contexts, and (4) studies situated in higher education, provided that they contained an explicit reference to school. The latter criterion was applied to retain studies conducted with university students that nonetheless addressed school-related contexts, for instance, studies on pre-service teacher education, on the use of digital textbooks in teacher training, or on the transition between school and higher education. Studies involving university students were excluded when they had no discernible reference to school teaching or teacher education and instead focused exclusively on the higher education context itself. Articles were further excluded if they (1) were not written in English or German, (2) did not relate to mathematics or geography (although studies that dealt with mathematics or geography in addition to other subjects were included), (3) were not accessible, or (4) did not explicitly deal with digital textbooks. The eligibility criteria were, however, specified before screening began and were applied unchanged throughout the process. The selection process was conducted and documented in accordance with the PRISMA 2020 statement (Page et al., 2021) and is shown in Figure 1.

The screening and selection of studies was carried out by the four authors, who are based at the two participating institutions. The records were divided at random into two halves, one of which was screened at the University of Münster and the other at the Free University of Bozen-Bolzano. Within each institution, titles and abstracts were screened by two researchers, so that every record was assessed by two people and screening decisions were reached by consensus rather than recorded as separate independent judgements. The same procedure was applied to the full-text assessment of the reports sought for retrieval. Records and reports whose eligibility remained unclear were discussed and decided jointly by one author from each institution. These two authors subsequently also carried out the coding of the included studies.

Data analysis

The data analysis was designed to systematically address the three research questions of this scoping review by structuring and aggregating the coded data derived from the final corpus of studies on digital textbooks in mathematics and geography. The analysis focuses on (a) the research foci and methodological approaches of the included studies (RQ1), and (b) the design features of digital textbooks explicitly addressed in the literature, with particular attention to accessibility-related aspects (RQ2 and RQ3).

The analytical procedure was based on a shared coding manual, developed by the University of Münster and applied consistently across all included studies. The analysis presented here draws exclusively on the final datasets (see Brnic et al., 2026).

Analytical framework and coding procedure

The coding manual specifies screening criteria, bibliographic variables, study characteristics, research foci, and a detailed set of design features of digital textbooks. The coding scheme was applied uniformly to studies from both subject domains, ensuring comparability between mathematics and geography.

For all design features, a binary coding scheme was applied (1 = feature explicitly addressed; 0 = feature not addressed). This coding reflects the presence or absence of an explicit treatment of a given feature in the article and does not involve any inferential interpretation. Cases initially coded as “unclear” were resolved through discussion within the research team and are not included in the aggregated analyses.

Each article was read in full prior to coding. Coding decisions were based solely on what was explicitly described, analyzed, recommended, or required in the article with regard to digital textbooks. Coding was conducted manually by the authors. To support orientation within the final included studies, the AI-assisted literature analysis tool NotebookLM (Google) was used supplementarily, operating exclusively on the uploaded primary studies. The coding manual and the full texts were uploaded to the tool, which was queried to locate passages relating to individual categories of the manual. The tool thus served for retrieval within the texts. It did not assign codes, and all coding decisions were verified and documented manually.

Coding for RQ1: research foci and research method

To address RQ1, each article was assigned to one primary research focus, representing the main content-related emphasis of the study. This categorization captures what the respective study primarily investigates in relation to digital textbooks. The research foci were not derived inductively from scratch but built on established classification frameworks for textbook and curriculum resource research. The starting point was the four-category framework proposed by Fan et al. (2013) for classifying mathematics textbook research, which was subsequently extended by Rezat (2024) through open coding procedures to develop more differentiated subcategories. Drawing on these frameworks, an initial category system was applied deductively to the included studies and then iteratively refined and specified in close interaction with the material, in order to adequately capture the foci of research on digital mathematics and geography textbooks. This combined deductive-inductive procedure resulted in the five research foci listed below. Analysis: systematic analyses of contents, structures, tasks, or properties of digital textbooks

Conceptual overview: theoretical or mapping contributions without a dominant empirical focus

Development and production: studies addressing the design, development, or production of digital textbooks or prototypes

Effects: investigations of the effects of digital textbooks on learning outcomes, achievement, motivation, engagement, or related variables

Utility and use: studies focusing on the use, implementation, or acceptance of digital textbooks in educational practice

Both the research focus and the research method were coded as mutually exclusive categories, based on the dominant orientation of the study.

Coding for RQ2: design features of digital textbooks

In this review, the term design features refers to the structural and instructional elements and functionalities that characterize a digital textbook, for instance its integrated visualizations, its feedback mechanisms, or its adaptive and data-based functions. The term is used in a broad, functional sense and does not denote graphic or layout design in the narrower aesthetic sense.

To address RQ2, a feature-based analysis was conducted. As with the research foci, the set of design features was not derived inductively but grounded in the existing literature on the characteristics of digital textbooks and digital curriculum resources. The starting point was the structural elements identified in earlier work on the features of digital (mathematics) textbooks, in particular the characteristics of digital curriculum materials described by Choppin et al. (2014) and the aspects of digital textbook presentation discussed by Usiskin (2018) and Leung et al. (2023). This initial, theory-based set of features was then specified and refined in interaction with the material in order to capture the design features relevant to the included studies.

For visual and representational features, the analysis followed the distinction proposed by Rolfes et al. (2020) between static, linear-dynamic, and interactive-dynamic visualizations: static visualizations do not change over time (e.g., images), linear-dynamic visualizations change over time without user control (e.g., videos, animations), and interactive-dynamic visualizations allow users to manipulate the representation themselves (e.g., by adjusting parameters).

The coded design features comprise: visual and representational features (static visualizations, linear-dynamic visualizations, interactive-dynamic visualizations);

interaction and feedback (interactive task formats, feedback on solution input);

adaptivity and data-based features (adaptivity, tracking or log files, AI-based functionalities);

navigation and support features (navigation tools, links or glossaries, personalization, collaboration);

textual elements (texts).

Coding for RQ3: accessibility

Accessibility was treated as a cross-cutting analytical dimension rather than as a separate research focus. Accordingly, it was coded as a dedicated category, independently of the research focus (RQ1) and the design features (RQ2), and is reported separately in order to illustrate how and in which analytical contexts accessibility is addressed across studies on digital textbooks in mathematics and geography.

For RQ3, each study was coded using a binary scheme (yes/no) according to whether accessibility-related aspects were explicitly addressed in relation to digital textbooks. A study was coded as yes when it explicitly addressed accessibility in one of two senses. The first sense is technical accessibility, that is, accessibility features in line with established standards such as the WCAG, including references to learners with disabilities, assistive technologies (e.g., screen readers, braille, text-to-speech), or accessible representations. The second sense is pedagogical-didactic accessibility, that is, the accessibility of content for diverse learners, including references to inclusive design frameworks such as UDL, cognitive accessibility, multimodality, or the adaptation of content to different learning needs. Studies that did not address accessibility in either sense were coded as no.

The two senses served as criteria for assigning a positive accessibility code, but they were not coded as separate variables. This decision follows the same binary logic applied to the design features under RQ2 and keeps accessibility comparable with the other coded dimensions. A fully two-dimensional coding, in which technical and pedagogical accessibility would have been assigned independently to each study, was not adopted because the two senses are frequently addressed jointly within a single passage, and a reliable separation between them requires a closer reading than the presence/absence judgement applied at the coding stage. The differentiation between the two senses that is reported in section 4.3 was developed in a subsequent qualitative reading of the 23 studies coded as addressing accessibility, in which the relevant passages were examined with regard to how accessibility was understood in each study. This qualitative step was carried out by the two researchers who had also coded the material, and the resulting reading was discussed within the research team. It is reported as an interpretive characterization of the studies rather than as a systematically coded and independently verified category.

Coding accessibility as a separate dimension makes it possible to determine not only how frequently accessibility is addressed in research on digital textbooks, but also in which analytical contexts, that is, in connection with which research foci and design features, it becomes relevant.

Inter-rater reliability

To assess the reliability of the coding, a subset of 45 studies was coded independently by two raters. Agreement was quantified using Cohen's kappa, computed separately for each coding dimension and interpreted following the benchmarks of Landis and Koch (1977). Agreement was almost perfect for the assignment of studies to research foci (κ = 0.91, 93.3% agreement) and for the coding of accessibility (κ = 0.87, 95.6% agreement). Agreement was substantial for the classification of the methodological approach (κ = 0.71, 77.8% agreement). For the design features, agreement was substantial overall, with a pooled value of κ = 0.67 (83.6% agreement across 225 paired judgements, computed on the common subset of 45 studies for which all relevant coding fields were available from both raters). The reliability check for the design features covers the first five features of the coding manual, computed on this same subset: static visualizations (κ = 0.59, 80.0% agreement), linear-dynamic visualizations (κ = 0.64, 82.2%), interactive-dynamic visualizations (κ = 0.55, 77.8%), feedback on solution input (κ = 0.73, 86.7%) and interactive task formats (κ = 0.82, 91.1%). These five features were selected because they were the only ones for which an independent double coding was available for a sufficient number of studies. They further correspond to the group of features identified in section 1.1 as particularly relevant to learning in the literature on digital textbooks, namely interactivity, visualization and feedback (see Berney and Bétrancourt, 2016; Castro-Alonso et al., 2019 Moreno and Mayer, 2007; van der Kleij et al., 2015). For the remaining design features, double coding was carried out only sporadically, since the coding of these features was continuously compared between the two institutions during the coding process. The resulting number of paired judgements is too small to support a meaningful reliability coefficient at the level of the individual feature. Although the pooled value indicates substantial agreement, the feature-level values range from moderate (interactive-dynamic visualizations) to almost perfect (interactive task formats), so that reliability is not uniform across the five features. Remaining discrepancies were resolved through discussion within the research team until consensus was reached.

Systematic research in Web of Science and Scopus, followed by a screening process in accordance with PRISMA-ScR guidelines, resulted in a final analysis corpus of 78 studies, 73 of which were in the field of mathematics and five in the field of geography (see Figure 1). The included studies were analyzed on the basis of the three research questions with regard to their research focus (RQ1), the design features of digital textbooks examined (RQ2) and the consideration of accessibility (RQ3). The results are presented in the following sections. A detailed overview of all included studies and their respective codings is available as a dataset on Zenodo (see Brnic et al., 2026). The search strategy underlying this selection is documented in Supplementary File S1.

Overall overview of the current state of research (RQ1)

The 78 studies included were assigned to five main areas of research: Development & Production, Utility & Use, Effects, Analysis and Conceptual Overview (see section 3.2.2). The distribution across these categories is shown in Figure 2. Table 1 breaks down the frequencies by subject.

With a share of 33%, Development & Production represents the largest research focus. Within this category, conceptual-theoretical work (n = 11) as well as qualitative studies (n = 7) and mixed-methods designs (n = 6) dominate. Purely quantitative approaches are hardly represented (n = 2). The second largest focus is Utility & Use, accounting for 28% of all included studies. Methodologically, this category is most strongly influenced by qualitative designs (n = 15), followed by quantitative (n = 4) and mixed-methods approaches (n = 3). Effects studies, which make statements about the relationship between the use of digital textbooks and learning variables, account for 17% of the corpus and are the only category that is predominantly quantitative (n = 10). The category Analysis accounts for 13% of the studies (n = 10), which belong exclusively to the field of mathematics and are predominantly qualitative in nature (n = 5). Conceptual Overview accounts for the smallest share and mainly comprises reviews (n = 4) and conceptual-theoretical contributions (n = 3). It is noteworthy that the latter two categories are exclusively occupied by mathematics publications. Table 2 provides a comprehensive overview of the research design.

With regard to the target groups examined, the analysis shows that school students are the most frequently studied group (n = 32), followed by teachers (n = 16). In a number of studies, both groups are examined together (n = 4), and in one case a community of interest, i.e., researchers and teachers together (n = 1, Bokhove et al., 2023). University students as an independent target group are represented in five studies. A substantial proportion of the studies do not have an empirical sample (n = 20), which points to the high proportion of conceptual-theoretical and design-oriented work in the corpus.

The temporal development of the research field, shown as a cumulative publication curve in Figure 3, shows a very low publication volume until around 2010. From 2015 onwards, a substantial increase can be seen, which accelerates particularly sharply in the years from 2020 onwards: 16 of the 78 studies included were published in 2024 and 2025 alone.

Overview of investigated design features of digital textbooks (RQ2)

To address RQ2, the coded data were aggregated to identify which design features of digital textbooks are most frequently examined in the existing literature on mathematics and geography education (see Figure 4). To account for the unequal number of studies per subject domain, frequencies are reported as the proportion of studies coding a given feature, both overall and separately for mathematics (n = 73) and geography (n = 5). Because the geography subsample comprises only five studies, geography values are additionally reported as counts in Figure 4, and the geography percentages should not be read as directly comparable with those for mathematics.

Across all included studies, the coded design features fall into three broad bands. A first group of core instructional and representational features is addressed in roughly two-thirds or more of all studies: texts (80.8%), interactive task formats (70.5%), static visualizations (69.2%), and interactive-dynamic visualizations (65.4%). This indicates that a substantial share of the literature concentrates on how content is presented and how learners interact with tasks in digital textbooks.

A second, intermediate group is coded in roughly half to two-thirds of the studies and comprises feedback on solution input (61.5%), navigation (61.5%), links or glossaries (60.3%), and linear-dynamic visualizations (57.7%). These figures point to usability, orientation, and visual-spatial representation as recurrently considered design aspects.

A third group of data-driven and system-adaptive features is addressed considerably less frequently. Personalization (43.6%) and adaptivity (37.2%) appear in a minority of studies, and collaboration (28.2%), tracking or log files (25.6%), and AI-based functionalities (9.0%) are coded least often of all features examined.

When disaggregating the results by subject domain, both mathematics and geography studies show a strong emphasis on textual elements, visualizations, and interactive tasks. In the geography subsample, textual elements and interactive task formats are addressed in all five studies, most further features in three or four of the five studies, and collaboration, tracking or log files and AI-based functionalities in none of them. Given that the geography subset comprises only five studies, however, these proportions are reported for descriptive purposes only and are not interpreted as a subject-specific comparison.

Overall, the results for RQ2 show that the included research most frequently addresses visible and interaction-oriented features of digital textbooks, while adaptive and analytic functionalities are coded considerably less often.

Accessibility in digital textbooks (RQ3)

Accessibility was analyzed as a cross-cutting dimension embedded within the feature-based coding scheme. A study was coded as addressing accessibility when accessibility-related aspects were explicitly mentioned, discussed, or operationalized in relation to digital textbooks.

Of the 78 studies in the final corpus, 23 address accessibility. This subset is unevenly distributed across the two subjects. Eighteen studies originate from mathematics education and five from geography education. This distribution should, however, be read against the disciplinary asymmetry already present in the corpus. While accessibility is addressed in only about one quarter of the mathematics studies, it is taken up in all five geography studies.

In line with the two-dimensional understanding of accessibility outlined in section 1.2, the 23 studies coded as addressing accessibility were examined qualitatively with regard to whether they frame accessibility in pedagogical or in technical terms (see section 3.2.4). Since this differentiation was not part of the binary coding scheme, the following observations characterize the studies qualitatively and are not reported as coded frequencies. Most commonly, accessibility is framed in pedagogical terms, for example through references to multimodal representations, the clarity of visualizations, or the support of diverse learning needs (e.g., Guo et al., 2017; Johansson et al., 2024; Putra et al., 2023). In these cases, accessibility is linked to general principles of instructional design rather than to formal technical standards. A subset of these pedagogically framed studies draws explicitly on Universal Design for Learning (UDL), which conceptualizes accessibility as a design principle of the learning offer through flexible representations and multiple pathways for engagement (e.g., Lee and Shin, 2023; Roschelle et al., 2012). Notably, the five geography studies all follow the pedagogical pattern, whereas technical framings are found exclusively in mathematics education.

Only a small number of studies explicitly refer to technical accessibility requirements, such as compatibility with assistive technologies or the braille representation of mathematical notation (e.g., Huang et al., 2015; Lewis et al., 2010 Yamaguchi et al., 2014);. An explicit reference to established guidelines such as the WCAG, by contrast, is rarely found in the corpus. Where technical accessibility is addressed, it is moreover directed almost exclusively at learners with visual impairments, for instance, through braille capability (Yamaguchi et al., 2014) or DAISY-based formats (Bouck et al., 2016; Huang et al., 2015), while other needs groups, such as learners with motor or hearing impairments, are largely absent. Across the included studies, accessibility is thus addressed through pedagogical framings far more often than through formal technical standards or assistive-technology criteria.

Studies addressing accessibility do not concentrate on a distinct set of “accessibility features” but engage with broadly the same range of design features as the corpus as a whole. Where differences emerge, they point towards features associated with flexible and learner-adaptive design. Among the studies dealing with personalization, almost half also address accessibility (16 of 34), and a comparable pattern is found for adaptivity (12 of 29). These proportions are higher than for more conventional features such as texts or static visualizations (see Supplementary File S2, Cross-tabulation of digital textbook features and accessibility coding).

In summary, the findings regarding RQ3 show that, whilst the topic of accessibility does feature in the studies examined, the qualitative reading of these studies indicates that it is unevenly distributed across the two perspectives. Pedagogical approaches predominate, whilst technical approaches are limited to a small group of mathematics studies focusing on visual impairments. Explicit references to formal standards or inclusive frameworks such as UDL remain rare.

The following discussion points address three areas of tension that emerge from the findings of the review. The first tension concerns the nature of the research field. It is young, growing rapidly and already exhibits clear areas of focus. However, on key issues such as the effectiveness or the quality of content in digital school textbooks, it is based on a comparatively narrow empirical evidence base. In addition, there is a distinct disciplinary asymmetry between mathematics education and geography education, which in itself constitutes a major finding.

A second tension lies between the conceptual expectations of digital textbooks and their actual research focus. Research focuses primarily on visible surface features (e.g., texts), whilst precisely those functions that conceptually distinguish digital textbooks from their printed counterparts and which are attributed high potential, such as formative assessment, personalization options or connectivity (Gueudet et al., 2018; Leung et al., 2023), have been comparatively little investigated.

A third tension concerns accessibility. Although accessibility aspects are addressed in some studies, they are rarely examined as a subject in their own right. In the majority of studies, accessibility is treated in terms of pedagogical accessibility, i.e., as a matter of learning support and design decisions in the sense of UDL (see CAST, 2024). Technical accessibility, which is based on established standards such as WCAG (World Wide Web Consortium, 2024) or explicitly refers to assistive technologies (e.g., Bouck et al., 2016; Huang et al., 2015 Yamaguchi et al., 2014);, is found, however, in only a small subset of the research. This finding requires explanation in light of the European educational policy context, which, among other things, highlights the need for the design of accessible materials through the European Accessibility Act (European Parliament and Council of the European Union, 2019).

Profile of a young field of research

The findings regarding RQ1 allow us to situate the field of research on digital textbooks in mathematics and geography (see section 4.1). They point to a field that is certainly vibrant, yet one whose focal points are clearly defined. At the same time, there is a clear asymmetry in subject-specific didactics between mathematics and geography.

The results can be understood as a reflection of a still young field of research. Together accounting for around 61%, studies from ‘Development & Production’ and ‘Utility & Use’ fall within those research priorities that typically dominate in the early phase of establishing new educational resources. Prototypes are designed and implementations tested (e.g., Gueudet et al., 2016; Guo et al., 2017 Hoch et al., 2018);, or acceptance and usage are observed (e.g., Sevimli, 2022; Wijaya et al., 2022).

In contrast, effects studies, the only category with predominantly quantitative designs, and systematic analyses are substantially less well represented in terms of numbers. Conclusions about the learning effectiveness of digital textbooks are therefore currently based on a comparatively narrow empirical basis, which suggests caution when interpreting assumptions about their impact. This distribution differs markedly from reviews of textbook research as a whole, in which content analyses traditionally account for the largest share (see Fan et al., 2013; Rezat, 2024).

The shift evident here corresponds to the observation that digital textbooks are often still in the trial phase, and consequently different research questions arise than with long-established printed works. At the same time, it complements the finding by Rezat (2024), according to which digital textbooks accounted for only around 16% of curriculum resource research in the period 2018–2023, to provide an interdisciplinary perspective on digital textbooks as an independent research subject.

One of the most striking findings of this review is the marked asymmetry between the two subject areas examined. In mathematics education, there is an established international research landscape focusing on school textbooks and, more recently, on curriculum resources in the broader sense, which is further consolidated by recurring review articles (see Fan et al., 2013, 2025; Rezat 2024; Rezat et al., 2021). Digital textbooks are increasingly emerging as a distinct sub-field within this research. In contrast, research into geography textbooks has traditionally focused more on content analyses of socially relevant topics and on evaluation criteria for teaching materials (Bagoly-Simó, 2021). Digital textbooks have so far been scarcely addressed systematically within this research tradition. In Bagoly-Simó’s (2021) overview of international geography textbook research from 1960 to 2020, digital textbooks are not dealt with separately. The asymmetry evident here is consistent with differences that have previously been described for the two research traditions, in which the digital has not yet been established as a distinct subject in geography education, either conceptually or empirically. It cannot, however, be attributed to the state of these research traditions alone. The restriction to two databases, the limitation to publications in English and German, and the very small geography corpus may equally have contributed to the observed imbalance, so that its magnitude should be interpreted with caution.

Research on the didactics of geographical education concerning digital textbooks currently appears to be at a considerably earlier stage of consolidation compared to research on the didactics of mathematics. The results bring the few existing studies in both subjects into common view, thereby creating an empirical foundation upon which further, specifically geographically oriented educational discussions of digital textbooks can build.

Between conceptual expectations and actual research focus

The findings on RQ2 can be interpreted as a tension between the conceptual expectations associated with digital textbooks and the actual focus of research on them (see section 4.2). Research concentrates on those features that are immediately visible on the user interface, e.g., texts, static and dynamic visualizations, interactive task formats, as well as feedback and navigation, whilst precisely those functions that conceptually distinguish digital textbooks from their printed counterparts, such as connectivity and adaptivity, are investigated comparatively rarely. This gives the impression that, in current research, digital textbooks are primarily conceived as interactive extensions of printed versions, supplemented by multimedia elements and task-based features.

One reason for this focus may be that it is directed at features whose effectiveness has already been demonstrated in numerous studies, for example the effectiveness of dynamic visualizations (Berney and Bétrancourt, 2016; Castro-Alonso et al., 2019), interactive digital tools (Hillmayr et al., 2020) and computer-based feedback (van der Kleij et al., 2015). While these features undoubtedly support engagement and understanding (e.g., Brnic et al., 2024; Wijaya et al. 2025),, the research focus thus remains largely confined to well-established, surface-level affordances.

Moreover, the frequent focus on navigation and linking structures underscores the importance of usability and orientation in complex digital learning materials. Effective digital textbook design must therefore integrate clear structural organization alongside rich content representations.

Nevertheless, analyses of school textbooks repeatedly show that these promising design features are often implemented only in a limited form in actual products. Interactivity is often limited to simple task formats, and task types differ only marginally from their printed counterparts (Behnke, 2021; Brnic and Greefrath, 2025; Rezat, 2025). Data-driven and system-level adaptive functionalities, tracking or log files, as well as AI-based functionalities, are substantially less frequently the subject of research than the aforementioned surface-level features. Taken together with the textbook analyses cited above, this indicates a gap between the expectations associated with digital textbooks and the attention that these design features receive in research.

This is particularly striking against the background of conceptual work, in which digital textbooks are defined precisely by their connectivity to other resources and actors (Gueudet et al., 2018). If digital textbooks are conceptually understood as part of a larger resource system but are empirically studied predominantly as isolated products, the potential of digital technologies for personalized learning support and for embedding textbooks within wider resource systems remains largely unrealized. Whether these distinctive features also prove effective in empirical terms is, for digital textbooks, still an open question. This opens up a clear opportunity for future research to move beyond surface-level affordances and to investigate adaptive, analytic and connective functions more systematically.

Accessibility as a research and design perspective

The findings on RQ3 show that accessibility, while present across the investigated studies, has not yet become an established research subject in its own right (see section 4.3). Although just under a third of the included studies touch upon accessibility, the analysis indicates that in most of them it figures as a secondary consideration rather than as the actual object of investigation. This pattern is worth interpreting more closely, since it shapes how accessibility is conceptualized in research on digital textbooks.

The two perspectives distinguished in the qualitative reading of the accessibility studies, pedagogical and technical accessibility, are addressed to markedly different degrees, and this imbalance is itself analytically significant. That the large majority of studies frame accessibility in pedagogical terms, embedded within general principles of instructional design, has a clear rationale. It situates accessibility within the question of how all learners can engage with the material and thus connects it to inclusive teaching. At the same time, it suggests that accessibility is predominantly understood as a matter of degree within good didactic design rather than as a distinct requirement. The technical perspective, by contrast, remains confined to a small group of studies and, within these, almost exclusively to learners with visual impairments. The near-absence of formal standards such as the WCAG and of needs groups beyond sensory disabilities indicates that, in most studies, accessibility is rarely operationalized as a systematic analytical category. Yet it is precisely such operationalization that would make accessibility verifiable as a property of the material itself.

A considerably smaller subset of studies does operationalize accessibility in this systematic sense, addressing it explicitly at the level of assistive technologies and established standards. That such work is sparse in the integrated studies does not mean it is absent from research more broadly. A dedicated review of assistive technologies for mathematics learning among visually impaired learners, for instance, documents an extensive and long-standing body of work on braille, audio-tactile and speech-based access (Shoaib et al., 2023). This research belongs to a different field, however. It develops largely in parallel to research on digital textbooks and is only rarely connected to it. Studies of this type follow a technical accessibility approach that also differs methodologically from pedagogical accessibility. Whereas the pedagogical perspective treats accessibility as something that emerges from how material is used in teaching, the technical perspective treats it as a verifiable property of the material itself, one that can be assessed independently of any particular instructional setting.

A particular missed opportunity emerges around UDL. UDL would, in principle, offer a conceptual bridge between the pedagogical and technical perspectives, as it links flexible representations and multiple pathways for engagement to accessibility as a design principle. It is notable in this context that personalization and adaptivity are associated with accessibility with above-average frequency, since both are themselves part of the repertoire of universal design. Yet the literature usually does not make this connection explicit, and UDL is invoked only sporadically across the studies examined. The integrative potential of such frameworks therefore appears largely unrealized.

Taken together, these observations point to a twofold need for future research. Accessibility should be addressed as an explicit object of investigation rather than emerging only as a by-product of research on usability or pedagogical quality. In addition, research should draw more systematically on integrative frameworks that connect the pedagogical and technical perspectives. The scarcity of such integrative work is increasingly being recognized in practice-oriented initiatives. Design guidelines developed within the same research context argue that established technical standards, while indispensable, are not sufficient on their own to capture the requirements of qualitative accessibility and inclusive teaching, and therefore call for an integration of didactic and accessibility considerations from the outset (Brnic et al., 2025). However, as long as the research community itself continues to treat the subject of accessibility as fragmented and conceptualized in an inconsistent manner, the field will lack the common analytical vocabulary required to assess the extent to which digital textbooks meet the accessibility requirements that now apply to digital products in accordance with the political regulations for digital products, in particular the European Accessibility Act (European Parliament and Council of the European Union, 2019).

Subject-specific implications for mathematics and geography

Subject-specific requirements regarding accessible design are only touched upon marginally in the literature. In mathematics, problem-solving environments involving symbolic notations, formulas and mathematical visualizations place particular demands on accessible representations, for example via accessible mathematical markup such as MathML or via Braille output. Although these requirements are discussed in a few technically oriented studies (e.g., Bouck and Satsangi, 2020; Yamaguchi et al., 2014), they are not systematically integrated into mainstream research on digital mathematics textbooks. In geography, the requirements are similarly high. For example, maps, spatial visualizations and layered geoinformation require textual alternatives, alternative navigation methods and carefully designed descriptions of visual content. However, there is no study that addresses these challenges in terms of technical accessibility; the five geography studies all remain at the level of pedagogical accessibility. In conjunction with the subject-specific didactic asymmetry discussed in 5.1, this results in a specific need.

Research perspectives and implications for practice

The findings outlined in the preceding sections point to several research gaps that can be consolidated into a coherent agenda for the further development of the field. This agenda concerns both research into digital textbooks and their design in specific development projects.

Strengthening the geography education perspective

The marked asymmetry between the two subject didactics, with mathematics accounting for the overwhelming majority of the investigated studies, should not be read in isolation. It partly reflects the generally smaller and less consolidated state of textbook research in geography education, where textbooks have long received only marginal scholarly attention and where research activity has been episodic rather than cumulative (Bagoly-Simó, 2021). Against this background, the more telling observation is not the difference in volume but the developmental stage of the field. Research on geography textbooks has remained largely content-oriented, whereas questions of design, production, and usage are underrepresented, and digitization in particular has been identified as a largely unexplored matter (Bagoly-Simó, 2021). This is consistent with the present review, in which the few geography studies do not yet establish digital geography textbooks as an independent empirical object.

The realistic implication is therefore not that geography education should match the research volume of mathematics education, but that it should develop digital textbooks as a distinct empirical subject in their own right. This includes systematic textbook analyses that delineate the specific characteristics of digital geography textbooks in comparison with their printed counterparts, as well as effectiveness studies that move beyond the conceptual and theoretical debate that has dominated to date (see Behnke, 2021; Guo et al., 2017). Closer methodological and conceptual integration with the more established research on mathematics learning materials could provide a productive starting point, without diluting the distinctive features of subject-specific didactics. Such development is particularly warranted given the continuing importance of textbooks in geography lessons (Bagoly-Simó, 2021).

Conceptual consolidation and empirical investigation of accessibility

The heterogeneity in the conceptualization of accessibility calls for clearer conceptual distinctions. Future research should distinguish more precisely between pedagogical and technical accessibility and systematically relate both interpretations, rather than pitting them against one another as alternatives. UDL offers a starting point for the pedagogical dimension, which would need to be linked to technical requirements, but it was explicitly drawn on as a framework in only a few studies in the present corpus (e.g., Lee and Shin, 2023). This includes the need for subject-specific operationalizations. For what accessibility concretely means in a digital mathematics textbook is not necessarily identical to what it must achieve in a digital geography textbook. Beyond conceptual clarification, there is a clear need for empirical studies that not only formulate accessibility requirements but also investigate how such features are actually implemented and used. This applies in particular to the perspective of learners with disabilities or specific learning needs, whose perspective has been represented only in exceptional cases in previous research on digital textbooks.

Broadening the focus on specific enhancement features

The focus on visible and surface-level features suggests that research should be specifically expanded to include those functions that conceptually distinguish digital textbooks from their printed counterparts: adaptivity, data-driven learning diagnostics, AI-based functionalities and connectivity-oriented functions. It is precisely at the intersection of these functions with inclusive learning environments that particular potential lies, but also serious risks, for example regarding data protection, algorithmic biases or the implicit normalization of certain learning trajectories. Research that reflects on and addresses these tensions is currently scarce and represents a key area of focus.

Methodological diversification

Finally, the profile of the research field described in 5.1 suggests a need for methodological diversification. Impact studies with controlled designs, systematic textbook analyses and triangulating studies that combine textbook analyses, usage studies and impact measurements would broaden the currently narrow empirical basis for claims regarding the learning effectiveness of digital textbooks. In addition, longitudinal studies that track implementation and usage processes over extended periods would be valuable. This is a methodological approach that, given the short history of digital textbooks, has so far scarcely been realized.

Limitations

When interpreting the present findings, several methodological and conceptual limitations must be taken into account. These are partly due to the nature of a scoping review as a mapping synthesis, and partly relate to decisions regarding the search strategy and coding.

A first limitation concerns the search strategy and the underlying databases. The search was limited to Web of Science and Scopus. Although both databases offer broad and quality-assured coverage of scientific publications, they do not index all relevant journals in the field of subject didactics. Studies published in non-indexed journals, in collections or as ‘grey literature’ are therefore underrepresented in the review. Furthermore, the language restriction to German and English may have contributed to a cultural and regional bias, as relevant research in other languages was excluded. In addition, it was not systematically recorded whether the included studies underwent a peer-review process, meaning that the methodological quality of the studies varies accordingly. A further limitation concerns the selection stage itself. Titles, abstracts and full texts were screened jointly by two researchers and decided by consensus rather than coded independently, so that no reliability coefficient can be reported for study selection.

A second limitation arises from the operationalization of the term ‘digital textbook’. The term is used inconsistently in the literature, and the boundary between digital textbooks, interactive learning environments and more comprehensive learning platforms is fluid. The review was based on a definition that, following Brnic et al. (2024) and Pepin et al. (2015), understands digital textbooks as technology-based curriculum resources with specific enhancement features. This operationalization was applied consistently but excludes studies that deal with similar materials under different terms.

Thirdly, the analysis is based on binary feature coding. This approach ensures a high degree of comparability between the studies and is appropriate for the mapping function of a scoping review. However, it disregards the depth and quality of the respective analysis. A study that merely mentions a feature in passing is given the same weight in the analysis as a study that examines it in depth. This limitation has a particular impact on RQ3, as, for example, the distinction between pedagogical and technical accessibility discussed in 5.3 emerged from a qualitative reading of the accessibility studies (see section 3.2.4) and is not reflected in the binary coding itself.

Fourthly, a key limitation concerns the comparison between subject domains. Due to the highly unequal sample sizes, quantitative comparisons between mathematics and geography are of limited reliability. With a geography sample of five studies, percentage distributions across research foci, methods or features are statistically insignificant and are heavily influenced by individual studies. Comparative statements between the subjects should therefore primarily be qualitative in nature, and percentage figures for the geography sample should be interpreted with corresponding caution.

Fifthly, the scoping nature of the review implies a fundamental limitation on the scope of its findings. The review aims to map and characterize the research field, not to assess the quality of evidence in individual studies (see Tricco et al., 2018). Statements regarding the effectiveness of digital textbooks or the quality of their designs are therefore limited to what the included studies themselves report. A systematic assessment of their methodological robustness was not part of this review.

The findings of this review therefore represent a snapshot of a rapidly evolving field of research. Individual statements may be relativized by new publications or developments in textbooks shortly after the search was completed. This applies in particular to areas characterized by high innovation dynamics, such as AI-supported functions and adaptive learning systems, whose current under-representation in research is likely to change in the coming years.

Conclusion

This scoping review mapped the research on digital textbooks in mathematics and geography along three dimensions: research foci, design features, and accessibility. The picture that emerges is that of a young and rapidly expanding, yet unevenly developed field. It is shaped by a marked disciplinary asymmetry, with mathematics education clearly dominating and geography education only beginning to address digital textbooks as a subject in its own right. At the level of design features, attention concentrates on immediately visible functions such as texts, visualizations, and interactive tasks, whereas the adaptive, data-driven, and connective functions that conceptually distinguish digital textbooks from their printed counterparts remain comparatively underexplored. Accessibility is present across the corpus but rarely constitutes an explicit object of investigation. It is conceptualized inconsistently, pedagogical framings clearly outweigh technical ones, and integrative frameworks such as UDL are invoked only sporadically.

Taken together, these findings indicate that research has so far engaged primarily with the more visible surface of digital textbooks, while their distinctive potential and their accessibility for diverse learners remain insufficiently understood. Against the background of current European accessibility legislation and the growing expectation that digital learning materials be accessible from the outset, the review underlines the need to establish geography as an independent research perspective, to investigate distinctive enhancement features more systematically, and to develop a shared conceptual vocabulary that connects pedagogical and technical accessibility. In this way, it provides an empirical foundation for future research and for the inclusive design of digital textbooks.