Accessibility within open educational resources and practices for disabled learners: a systematic literature review

The number of disabled students is rapidly increasing worldwide, but many schools and universities have failed to keep up with their learning needs. Consequently, large numbers of disabled students are dropping out of school or university. Open Educational Resources (OER) and Open Educational Practices (OEP) contain several relevant features, including the possibility of reusing and remixing, which have led researchers to consider using OER and OEP to facilitate meeting the needs of disabled and functional-diverse students in order to increase their accessibility and e-inclusion capabilities in educational settings. The very limited research to date, however, has provided a limited holistic understanding of accessibility within OER and OEP in order to aid researchers in pursuing future directions in this field. Therefore, this paper systematically reviewed 31 papers to provide insights about functional diversity within OER and OEP. The results obtained highlighted that accessibility is still in its infancy within OER and that researchers should focus more on considering the four accessibility principles — perceivable, operable, understandable and robust — when providing OER. Additionally, while several researchers have focused on several issues related to accessibility within OER, limited focus has been given to assistive technologies using OER. Finally, this paper provides several recommendations to increase accessibility within OER and help design more accessible OER for students with functional diversity.

Open Educational Resources (OER), defined as 'teaching, learning and research materials in any medium that may be composed of copyrightable materials released under an open license, materials not protected by copyright, materials for which copyright protection has expired, or a combination of the foregoing' (UNESCO, forthcoming), have the potential to contribute to reaching this objective by increasing access to learning as well as improving the quality of the learning experience (Ehlers, 2011). The OER movement is based on the idea that educational resources (e.g., content or course designs) should be released under licenses that allow anyone to freely access, retain (e.g., download, duplicate, store), reuse, revise (e.g., translate, adapt, modify), combine and-or re-share them (Tlili, Huang, Chang, Nascimbeni & Burgos, 2019). The use of OER for teaching in an innovative and collaborative environment is referred to as Open Educational Practices (OEP). Ehlers (2011), p. 4 defined OEP as 'practices which support the (re)use and production of Open Educational Resources through institutional policies, promote innovative pedagogical models, and respect and empower learners as co-producers on their lifelong learning paths'. Research is coalescing around the fact that these practices can help enhance learning quality, access and effectiveness in universities (Weller, 2014).
Despite the growing number of OER (Hoosen & Butcher, 2019) and the policy attention devoted to OER accessibility, as demonstrated by the presence of guidelines to increase the accessibility of OER within the Ljubljana OER Action Plan (UNESCO, 2017), the extent to which OER are actually accessible is currently being questioned. Accessibility refers to the use of a product, service, framework or resource in an efficient, effective and satisfying way by people with different abilities (ISO 9241-171, 2008). Functional diversity is a key issue in the development of any online resource, including OER, since it is potentially focused on almost every single user. The approach has moved from handicapped users (essentially, those with motor, cognitive or sensorial impairments) through accessibility (improving specific issues to facilitate a better user experience) to functional diversity and e-inclusion (of any feature of any user who requires additional support, like the ones associated with elderly or those on sick leave) (Iniesto, Covadonga, & Moreira Teixeira, 2014;Sanchez-Gordon & Luján-Mora, 2013;Tekleab, Karaca, Quigley, & Tsang, 2016).
The present paper aims to provide a holistic and systematic review of the literature in the field of the accessibility and functional diversity of OER and OEP, as a valuable guide for better designing open educational ecosystems that support inclusive learning, improving the potential effect of OER on twenty-first century teaching and learning for learners with different needs. This is particularly urgent since recent data estimates that 15% of world populationmore than a billion peoplelive with some form of disability (World Health Organization and World Bank, 2011). The structure of the paper is as follows. Section 2 presents the background of the research, section 3 details the research method, section 4 presents and discusses the obtained results, and section 5 concludes the paper with a summary of the findings, limitations and potential future directions.

Background
According to the World Health Organization, disability cover[s] impairments, activity limitations, and participation restrictions. An impairment is a problem in body function or structure; an activity limitation is a difficulty encountered by an individual in executing a task or action; while a participation restriction is a problem experienced by an individual in involvement in life situations. (World Health Organization, 2015).
The Office for Civil Rights (OCR) of the U.S. Department of Education defines 'accessible' as meaning that a person with a disability is afforded the opportunity to acquire the same information, engage in the same interactions, and enjoy the same services as a person without a disability in an equally effective and equally integrated manner, with substantially equivalent ease of use.
In educational contexts, accessibility for disabled students means that, in order for all to have equitable learning experiences, the learning experience, including its learning content and teaching process, should be adjusted according to students' needs, including their disabilities. While people with disabilities have the same educational needs as others, they are less likely to attend schools and graduate, and consequently may face difficulties in finding jobs in future (Ingram, 1971;Iwarsson & Ståhl, 2003;World Health Organization and World Bank, 2011). Various international policies, including the United Nations 2030 Agenda for Sustainable Development (United Nations, 2015) and the UNESCO Education for All initiative (UNESCO, 1990), have highlighted the importance of providing fair learning experiences for all students regardless of their differences. Still, a great proportion of schools and universities fail to properly address equitable access, especially with regard to disabled students (Catlin & Blamires, 2019), partly due to the lack of effective teaching methods and content targeted to these student categories (Virnes, 2008).
In the area of web accessibility, several standards released by the Web Accessibility Initiative (WAI) of the World Wide Web Consortium (W3) can be applied to OER. Among these standards, WCAG 2.0 has been widely accepted and adopted (W3C., 2012) and is based on four attributes that lay the necessary foundations for anyone to access and use websites, as shown in Table 1. Based on these four attributes, 12 guidelines and 61 success criteria are provided, categorised into three levels of conformance: AAA (highest), AA or A (lowest) (Crespo, Espada, & Burgos, 2016;W3, 2008). Table 1 shows that OER can increase the accessibility of web-based education in many ways. This potential is mainly connected to the inner OER features of re-using, remixing and redistributing learning content that can help adapt existing materials to disabled students without having to develop resources from scratch. OER can serve the needs of those with diverse abilities for a number of complementary reasons: Permissions granted by an open license remove legal barriers to adapting and customising OER, making it possible to create learning environments that are more flexible and robust for all students. OER offer the opportunity for instructors to curate materials authored by a diverse set of individuals, including those who identify as disabled, normalising and reducing stigma while sharing viewpoints that have historically been marginalised. Unlike commercially published materials, OER that are adapted to meet accessibility requirements can be retained and freely shared with communities, reducing duplicative work at and across institutions. OER adoption can reduce costs, which benefits all students but can be especially beneficial for students with disabilities who may face additional financial pressures.
It is more common for OER to be shared in formats that can be adapted for accessibility, unlike proprietary publisher content, from whom editable files are notably difficult to obtain (Thomas, 2018). Hejer, Khribi, and Jemni (2017) mentioned that despite the fact that the OER paradigm can facilitate inclusive learning by reusing the open resources in a way which caters to the needs of disabled students, limited work has been done to achieve this purpose. Similarly, Iniesto, McAndrew, Minocha, and Coughlan (2017) stated that few Massive Open Online Courses (MOOCs) are fully accessible for disabled students. Undeniably, not enough research is being conducted to support inclusive and equitable learning using OER (Navarrete, Peñafiel, Tenemaza, & Luján-Mora, 2019). Specifically, to our knowledge, only one conference paper has conducted a systematic literature review to investigate the actual accessibility of OER for disabled learners (Moreno, Caro, & Cabedo, 2018), providing only information about the trends of OER and accessibility without summarising and discussing findings related to accessible learning within OER and OEP. In addition, while several literature reviews have been conducted to better understand the use of OER for the general student population, no literature review has focused on investigating the work done on the accessibility of OER and OEP. To fill this gap, this paper presents a systematic literature review to understand how the application of OER and OEP can increase learning accessibility.

Text Alternatives
Provide a variety of forms that people need for non-textual content, such as large print, Braille, and so on.

Time-based Media
Provide access to time-based media.

Adaptable
Ensure that all OER are available in some way to all users.
Distinguishable Make the default presentation easy to perceive by people with disabilities.
Operable OER, including the content and interface, must be operable for users.

Keyboard Accessible
Make all functionalities achievable by using the keyboard.
Enough Time Provide enough time for users to use OER.

Seizures
Do not design OER in a way that might trigger seizures.
Navigable Support navigation and retrieval functions.
Understandable OER, including the content and interface, must be understandable by users.

Readable
Make OER text readable and understandable.

Predictable
Make OER contents display and operate predictably.

Input Assistance
Provide more assistance to avoid and correct mistakes.

Robust
OER must be robust enough that it can be accessed by a variety of types of user agents, including assistive technologies.

Compatible
Increase compatibility with the current and future user agents, especially assistive technologies: i.e., screen reader or Braille display devices.

Methodology
A rigorous literature review is an important step that builds the foundation for knowledge accumulation, which in turn facilitates the expansions and improvements of theories, closes existing gaps in research and uncovers areas previous research has missed (Marangunić & Granić, 2015). This study presents a systematic review based on published papers related to OER and OEP for learning accessibility, with particular reference to disabled students. It follows the steps reported by Okoli and Schabram (2010) as described in the next subsequent sections.

Investigated research questions
To gain insight into the use of OER and OEP for accessible learning, a systematic review is needed. Specifically, this study attempts to answer the following research questions: RQ1. What are the trends in publications on learning accessibility using OER and OEP in terms of time series, country and keyword distribution? RQ2. What kinds of disabilities and issues were investigated in the identified papers? RQ3. Which assessment methodologies were used in the identified papers?

Search strategy and inclusion/exclusion criteria
To answer the above research questions, several keywords were adopted as follows: accessib* AND Open AND Educational Resource*, accessib* AND OER, accessib* AND Open Educational Resource, accessib* AND OEP, accessib* AND Open Pedagogy, accessib* AND Open teaching, accessib* AND Open assessment, accessib* AND Open educational Practices, Inclusive learning AND Open educational resource, Inclusive learning AND OER. The search was conducted in several databases, including ScienceDirect, Wiley Online Library, IEEE Xplore Digital Library, Core Collections of Web of Science and Taylor & Francis Online. ResearchGate, a network for researchers to share, discover and discuss research, was also used to retrieve the related papers. The obtained papers were then filtered based on specific inclusion/exclusion criteria. Specifically, we excluded papers that: (1) were not in English; (2) did not discuss openness using OER and OEP for learning accessibility; (3) did not focus on disabled students; or (4) did not have available full-text online. A total of thirty-one papers were finally included during the review process. Figure 1 presents the selection procedure of papers during this review process.

Data extraction and analysis
Each study was then reviewed and examined based on seven items, as presented in Table 2. These items provide information to answer the above research questions and conduct the synthesis. Finally, a qualitative synthesis was conducted to answer the research questions.

Results and discussion
Trends in publications on learning accessibility using OER and OEP

Distribution by year
As shown in Fig  accounted for more than 60% of all the production of the last decade. Additionally, the year 2016 saw a peak in interest in this area, probably connected with the fact that the UN 2030 Agenda for Sustainable Development was launched in 2015, providing an impetus for research in the areas of accessibility and inclusion.

Distribution by country
The distribution of the first author's countries is presented in Fig. 3, showing that authors from only nine countries have led research about OER and OEP for accessible learning. This shows that the use of OER and OEP for inclusive learning is still in its infancy and that more awareness should be raised to encourage further investigation in this field. In particular, authors from Ecuador had 11 papers related to this topic, accounting for more than one third of all papers, followed by Spain, with six papers. Ecuador is indeed considered as a leading country in the field of disability support, since the government proposed in 2007 several policies to address the needs, including educational needs, of disabled persons. Spain has long attached great importance to inclusive education; as early as 1982, Spain passed legislation to integrate disabled youth in schools. In 1985 the decree on special education moved many disabled children from special schools to mainstream schools. In 1994, the United Nations World Conference on Special Needs Education was held in Spain, where the fundamental principle of inclusion at school was declared and widely endorsed. Interestingly, four out of the nine countries present at that conference (Ireland, Italy, Spain and the UK) have adopted the Web Content Accessibility Guidelines 2.0 (WCAG 2.0) noted earlier (W3, 2017).

Distribution by keyword
Finally, the keyword distribution of the 31 research papers in the systematic review was analysed in order to understand the use of OER and OEP for accessible learning more deeply. Keywords with similar meanings, such as 'OER' and 'Open Educational Resources' or 'Learning object' and 'LO', were normalised. The final distribution of the keywords is presented in Fig. 4. It can be seen that accessibility, OER and disability are the most commonly used keywords in the 31 papers reviewed. In particular, disability and accessibility focus on the category of students on which these research papers focus, while OER focus on the category of education that can contribute to improving the accessibility of earning opportunities. Importantly, we discovered that the term Open Educational Practices (OEP), as well as sub-terms, such as open pedagogy, open teaching and open assessment, have not yet been discussed in the literature when it comes to accessible learning. Therefore, in the subsequent analysis we will focus only on accessibility and OER.

Disabilities and issues investigated
As shown in Table 3, when investigating the use of OER, researchers focused on several disabilities, including visual disabilities, hearing disabilities, motor disabilities, speech disabilities, cognitive disabilities and aging-connected disabilities. Researchers paid almost Fig. 4 Distribution of keywords in the reviewed research papers equal attention to different types of disability, including seven studies on visual disabilities and hearing disabilities, respectively, and six papers on motor disabilities and cognitive disabilities. It is obvious that aging also imposes certain limitations on the ability of humans, so researchers have also considered it. It should be noted that some papers discussed more than one disability. For instance, Zervas et al. (2014) developed an online teaching and learning portal for students with visual and/or hearing disabilities. The use of OER to address the above disabilities was discussed from five different angles: system design, personalisation, metadata, authoring tools and OER accessibility framework/architecture. As shown in Table 4, most authors focused on system design to increase accessibility and usability for students with disabilities. For instance, Ngubane-Mokiwa (2016) conducted a literature review and identified several guidelines to facilitate MOOC access for visually impaired students. These guidelines are from three different perspectives: (1) multiple means of representation, which focuses on the strategies to make MOOCs accessible; (2) multiple means of action and expression, which focuses on the strategies that facilitates user actions on MOOCs; and, (3) multiple means of engagement, which focuses on strategies to provide accessible interaction within MOOCs.
Several researchers also analysed personalised learning experiences based on the 'type of disability' or 'user profile' as a personalisation parameter. For instance, Zervas et al. (2014) designed an OER-based educational portal to facilitate learning and teaching for students with different disabilities, including those with visually impairments. Similarly, Navarrete and Luján-Mora (2018) developed an OER website that takes into consideration the disability of students, including visual and hearing disabilities, as a personalisation parameter. This 'disability-personalisation' path is extremely relevant, as recognised by the National Academy of Engineering, which mentioned that personalised learning is one of the fourteen most important challenges of the twenty-first xentury . Other researchers focused on discussing metadata, defined and machine-processable data that describe resources, either digital or nondigital (Haslhofer & Klas, 2010), in inclusive learning using OER and OEP. An accurate metadata set can enhance the retrieval of educational resources and provide a friendly navigation experience. For instance, in order to better describe and identify resources, Navarrete and Luján-Mora (2018) applied a subset of descriptors from the Learning Object Metadata (LOM) standards. Similarly, Navarrete and Luján-Mora (2014) applied other metadata standards, including DCMI (Dublin Core Metadata initiative) and AfA (Access for All). Some researchers have put forward innovative frameworks to improve the accessibility of OER.  argued that the development of a framework for improving web accessibility should be based on existing standards, such as WCAG 2.0, and proposed a framework for enhancing the accessibility and usability of open courseware sites. Innovative architectures are also presented by Sanchez-Gordon and Luján-Mora (2016) as ways to improve the accessibility of MOOCs and OER.
Finally, some researchers have focused on developing authoring tools for accessible OER. For instance, Mulwa et al. (2016) developed an OER authoring tool to facilitate the creation of OER for students with visual disabilities by selecting the navigation methods and text sizes. As shown in Table 4, only two papers focused on authoring tools to develop accessible OER. This might explain the limited number of fully accessible OER. Therefore, more focus should be put on developing tools that can help educators create and publish OER for disabled students. Additionally, no reviewed paper discussed the accessibility of OER from the assistive technology perspective. Given that different assistive technologies for disabled persons exist within different Operating Systems (OS), OER designers should try to make their resources compatible with as many assistive technologies and OS as possible in order to ensure high accessibility.

Assessment methodologies used
Based on the review of the 31 identified studies, 16 papers conducted assessments to evaluate the accessibility of OER, while the 15 remaining papers did not conduct any assessment. Specifically, to assess the accessibility of OER, three different methods were used, as shown in  Iniesto & Rodrigo, 2014;Navarrete & Luján-Mora, 2015a), enabling the system to better understand the problems and requirements of people with impairments. For instance, the simulator named  Sanchez-Gordon & Luján-Mora, 2016 No assessment N/A Coughlan et al., 2016;Hejer et al., 2017;Iniesto and Rodrigo, 2018;Iniesto & Rodrigo, 2016;Iniesto et al., 2017;Iniesto et al., 2019;Kourbetis & Boukouras, 2014;Kourbetis et al., 2016;Morales and Benedi, 2017;Moreno et al., 2018;Navarrete et al., 2016;Politis et al., 2014;Sanchez-Gordon & Luján-Mora, 2015;Yalcinalp & Emiroglu, 2012;Zervas et al., 2014 aDesigner, used by Iniesto and Rodrigo (2014) and , aimed to simulate the use by people with visual disabilities in order to help the designer assess the extent to which a given content is accessible to users with that particular disability. Finally, manual assessment is mostly based on users' questionnaires (Avila Garzon, 2018;Avila Garzon et al., 2016;Caruso & Ferlino, 2009;Mulwa et al., 2016;Navarrete et al., 2019;Navarrete & Luján-Mora, 2015a;Navarrete & Luján-Mora, 2018;Sanchez-Gordon & Luján-Mora, 2016). In these cases, the purpose of the questionnaire is to obtain a qualitative analysis to appreciate the users' experience of the process of using a given OER (Navarrete et al., 2019), based on questions like 'Is it easy to learn how to use the website?' or 'Can the user resolve the tasks on the website without unnecessary effort?' (Navarrete & Luján-Mora, 2018). Several researchers, however, claimed that using questionnaires may not be motivating for learners, since they are typically too long. Additionally, learners may not fully reveal their experiences and may try to respond optimistically when they feel that they are being assessed by others (Okada & Oltmanns, 2009). To counterbalance these attitudes, given the rapid growth of technology and the era of big data and learning analytics, researchers should focus more on using the data generated by learners to obtain insights about the accessibility of OER-based learning processes. If we consider that the accessibility of OER and OEP should aim at enabling all users, including disabled ones, having equitable learning opportunities, this focus on technical accessibility suggests that the research on OER and OEP for disabled learners is still in its infancy, since most researchers have focused on a rather superficial analysis that does not rely on rich datasets. Therefore, further research should be conducted to investigate how OER and OEP facilitate the deployment of accessible and inclusive learning from a more holistic perspective. WCAG 2.0 provides guidelines on how to make web content more accessible to people with disabilities and four principles to lay the foundation of Web accessibility (W3, 2008). Table 6 presents the results of the review along with the four accessibility attributes presented in the Background section: perceivable, operable, understandable and robust. It appears that the majority of researchers discussed accessibility as one concept without considering specific accessibility attributes. Table 6 shows that the general OER accessibility level could be improved: among the 16 papers which reported accessibility assessment results, 15 generally agreed that there was much room for improvement in the accessibility of OER, especially for disabled users. For instance, the accessibility evaluation results by Iniesto and Rodrigo (2014) show a low degree of compliance of the analysed OER with the WCAG 2.0 accessibility guidelines. Navarrete et al. (2019) also conclude that neither the OER website interface nor the educational resources are fully accessible.
If we analyse the accessibility attributes individually,  stated that more errors are found under the attributes 'robust' and 'perceivable', which account for 50% and 31.81%, respectively, of the errors made when using the automatic tool TAW. On the other hand, for the attributes 'operable' and 'understandable', the percentage of errors is 20% and 17.64%, respectively. After accessibility evaluation with TAW of four OER platforms, including MERLOT, OCW UPM, OER COMMONS and OLI, similar results were reported in Navarrete and Luján-Mora (2015c), which showed that the greatest number of warnings are annotated under the attributes 'robust' and 'perceivable', while all of these warnings may be related to some issues that need to be

Conclusion, recommendations and future directions
This study presented a systematic review of the use of OER and OEP to provide accessible learning. The final notes based on the results discussed above (in the three presented research questions) are as follows: A limited number of countries (nine) were involved in the investigation of the use of OER and OEP for accessible learning (research question 1). Therefore, researchers worldwide should be encouraged to get involved in this research field. This can be changed by raising awareness about the new opportunities that OER and OEP could provide to disabled students for effective accessible learning, or by launching new projects or policies (e.g., governmental or institutional) that encourage the use of OER and OEP for inclusive learning.
Only two papers discussed the development of authoring tools with features to create accessible content, which might explain the reasons for having limited online OER and OEP for disabled students (research question 2). This should be changed by developing more inclusive authoring tools (that work with different functional diversities) that educators can use to create and publish open content.
Most assessments conducted focused only on the accessibility of the provided OER (research question 3). Therefore, more research should also be conducted to investigate the effectiveness of OER and OEP in providing accessible learning experiences and enhancing disabled students' learning achievements. There is still much room for improvement in OER accessibility (research question 3). Therefore, researchers and practitioners should consider different accessibility guidelines (e.g., WCAG 2.0) while developing their OER platforms, tools and devices. This helps provide an effective approach to accessibility, functional diversity and einclusion in educational settings. Only three assessment methods are used: automatic tools, simulator tools and manual tools (research question 3). Therefore, in the era of big data, researchers and practitioners should also begin applying learning analytics for more accurate assessment of the accessible learning experience provided to disabled and functional-impaired students. Among the four accessibility attributes, 'robust' has the highest percentage of errors (research question 3). Therefore, OER developers should place more emphasis on OER's compatibility with most assistive devices, as well as operating systems (Windows, Mac OSX and Linux).
In addition, the authors consider direct support to educators a key issue, so that they learn the foundations of functional diversity, develop the skill set to operate learning resources under these terms and are fully aware of the significance of and need for specific actions around the topic. Indeed, providing specific competencies and training for educators are a challenge but nonetheless a required measure to improve the impact of functional diversity and accessibility on the educational system.