Growth and Collaboration in Massive Open Online Courses: A Bibliometric Analysis

Massive open online courses (MOOCs) are an important approach for achieving UNESCO’s aim of open and accessible education. However, there are concerns regarding fragmentation or bias of MOOCs toward certain disciplines or countries. This study sought to: (a) examine how MOOCs research has evolved and is distributed, (b) determine what key areas are discussed in MOOCs research, and (c) identify the major players in MOOCs research and their collaborations. This study conducted a bibliometric analysis of 3,118 scholarly works related to MOOCs as recorded in the Scopus database in July, 2019. Specifically, we analyzed the evolution of MOOCs research by examining (a) published studies, (b) source titles, (c) types of sources and documents, as well as (d) the languages in which the documents were written in. We further analyzed the key areas of MOOCs research by looking into common subject areas, keywords used most often, and title analysis. Finally, we sought to increase our understanding of the major players in MOOCs research and their collaborations by examining (a) which countries contributed most to MOOCs research, (b) the main institutions involved, as well as (c) authorship and citation analysis. Findings indicated that in their early development starting in 2009, MOOCs caught the attention of scholars from both the East and the West, and the number of publications grew consistently over the 10 years after that. MOOCs research has been well distributed but has yet to adequately encourage inclusiveness. There has been healthy cross-country collaboration, but there is a gap in MOOCs research originating from certain countries as compared to the rest of the world. Our findings provide important input towards improving the inclusivity and global reach of MOOCs.


Introduction
Education is a human right, and massive open online courses (MOOCs) are an important tool whereby digital technology may be used to enhance access to quality education and lifelong learning opportunities for all. The number of MOOC offerings has grown exponentially, partly due to the Internet revolution as well as in response to the call to address the need for access to quality education in an equitable and affordable manner as inspired by the United Nations Sustainable Development Goal 4, which forms part of a universal agenda (United Nations General Assembly, 2015). In the period between 2012 to 2013, MOOCs came to be widely accepted by universities around the world, and outsourcing companies were launched to provide the necessary infrastructure (Baggaley, 2013). Since then, MOOCs have been a popular research topic-rapidly developing, while inspiring new approaches, innovations, assessments, and discussions.
Several studies have looked into trends in MOOCs research. Review studies have not only focused on different time periods, but have also examined different research goals, perspectives, and contributions. Liyanagunawardena, Adams, and Williams (2013)  Several studies have looked into the interdisciplinary aspects of MOOCs research. Studies revealed common research themes (Ebben & Murphy, 2014;Gašević, Kovanović, Joksimović, & Siemens, 2014), as well as research methods used and dominance of researchers from the field of education (Gašević et al., 2014). Gašević et al. (2014) also raised the concern of fragmentation in the research community and the need to enhance interdisciplinary efforts, but their study focused only on proposals submitted to the MOOC Research Initiative. Similarly, Veletsianos and Shepherdson (2015) found that MOOCs research published from 2013 to 2015 was mostly conducted by researchers from the education and computer science disciplines, though an interdisciplinary trend was also emerging. Veletsianos and Shepherdson (2016) further examined the geographic distribution, publication outlets, citations, data collection, and analysis methods of research focusing on MOOCs during 2013 to 2015. This study, however, excluded literature authored in languages other than English, and recommended that future research examine whether MOOC literature was biased towards certain countries or regions.
Other aspects of MOOCs research have also been studied. Deng and Benckendorff (2017) indicated that most research has used surveys, interviews, and logged files to understand instructors' and students' use of MOOCs. Ichimura and Suzuki (2017) analyzed literature focusing on MOOC course design. Zancanaro and Domingues (2017) (Zhu, Sari, & Bonk, 2018). However, these studies looked at only a small amount of MOOC literature, which did not show the bigger picture of MOOCs as a global movement.
Previous studies have focused on understanding MOOCs from various perspectives, but little has been done to determine whether MOOC development is equally shared or collaborated on in different parts of the world. This question is crucial, since MOOCs are viewed as a tool to reduce the educational gap across the world. As MOOCs require technology infrastructure, digital skills, and language fluency, these factors could also potentially increase the digital divide (Jiang, Williams, Warschauer, He, & O'Dowd, 2014;Lee, Hong, & Hwang, 2018) and cause serious social polarization across the world. There are also concerns about the often-overlooked cultural dimension of MOOC providers offering global education solutions (Nordin & Norman, 2018). • subject area; • frequency of keywords; and • title analysis (e.g., frequency of words and phrases). c) Major players and research collaboration: • countries with most contributions; • main institutions; • authorship analysis; and • citation analysis.
The purpose of this study was to gain a more in-depth understanding of the MOOC phenomenon, particularly with respect to its global reach and collaborations. It was necessary to examine the latest data in order to help researchers propose recommendations for future research in the development of

Method
This bibliometric study accessed the Scopus scientific database to analyze publications with the word MOOC or massive open online learning in their title. It considered all types of documents published in the Scopus database from the year 2009 until 2020. Scopus is one of the largest abstract and citation database of peer-reviewed literature; it contains approximately 23,700 peer-reviewed journals as well as over 24,000 titles, 360 trade publications, 750 book series, 195,000 non-serial books, and 60 million records from various areas of knowledge. Such a large database is able to provide a comprehensive overview of the world's research output. Scopus is also recognized by the international scientific community as one of the main sources of relevant information.
This study employed bibliometric analysis and used quantitative and statistical analysis to describe distribution patterns of research articles within specific topics and time periods (Martí-Parreño, Méndez-Ibáñez, & Alonso-Arroyo, 2016 We analyzed the results in various ways in order to provide input in response to our research questions.
Several results were directly retrieved from Scopus through the analyze search results function. Other results were inserted manually or exported to a new Excel file. From the file created for all the results, information such as percentages was analyzed. We also used VOSviewer to generate images to help with data interpretation. After the results were identified, analyzed, and synthesized, we wrote up the final report, which presented the findings and analysis. Through this paper we hope to contribute meaningful insights on the trends apparent in publications on MOOCs. Researchers can use these findings as a basis for future studies and discussions to enrich and further develop this area of research.

Results
This section deals with the results obtained from the bibliometric analysis related to the following

Evolution and Dissemination of MOOCs Research
To address the question of the evolution of MOOCs research and trends in its dissemination, this study analyzed the following data: (a) number of publications by year, (b) source title, (c) source and document type, and (d) document's language.
Publications by year. Table 1  publications in 2019, this study was conducted just past the midway point of July, 2019. Thus, the full number of documents for the year were yet to be published. In contrast, some journals had already produced their 2020 publications, so these numbers were also recorded by the Scopus database. Sources and document types. This study also sought to determine where MOOCs documents had been published by analyzing the data based on document source types. Table 2 shows that journals were the most common source, representing 1,322 (42.40%) of the total, followed by conference proceedings (n = 1,199; 38.45%) with a barely 4% difference only. Trade publications, normally intended for a specific industry, trade, or type of business and usually published in the form of a magazine periodical with the topical subject, were the least frequent document type (n = 11; 0.35%).
Although these trade publications were seldom referred to, they are also scientifically relevant and useful in influencing policies on MOOC implementations. The data were also analyzed based on document types. The Scopus database focuses on primary document types from serial publications, which means that the author is also the researcher in charge of the presented findings. Secondary document types, where the author is different from the person conducting the research, such as book reviews, are not included in Scopus document types. As a result, our analysis revealed the volume of researchers conducting research on MOOCs and their publications.
As shown in Table 3, nearly half of the total publications came from documents presented at a conference or symposium (n = 1,518; 48.69%). This was followed by articles of original research or opinion (n = 1,146; 36.75%). Book chapters represented 8.11% (n = 253) of the publications on MOOCs.
The other types of documents, such as reviews, editorials, letters, notes, books, conference reviews, short surveys, and erratum, each represented less than 2% of the total publications, respectively.

Key Areas of MOOCs Research
The key areas of MOOCs research were analyzed in terms of (a) main subject areas, (b) frequency of keywords, and (c) document titles.
Subject area. This study classified the documents based on their subject area, as presented in Table 6. The data showed that research on MOOCs has emerged in a variety of subject areas. Nearly 60% of studies involving MOOCs were in the area of computer science, representing 59.33% (n = 1,850) of the total articles, followed by a significant number of publications in the social sciences (n = 1,711; 54.87%). The subject areas of engineering, mathematics, decision sciences, business, management and accounting, arts and humanities, and medicine each accounted for more than 100 documents on MOOCs.  (Sweileh et al., 2017).  Table 7 summarizes the most frequently used keywords in MOOCs studies. After excluding the core keywords related to the search query, the data further revealed that e-learning was the keyword most associated with MOOCs (n = 1,031). The keyword "massive open online course" was always used either in full or as an abbreviation (MOOC) and was also presented interchangeably either as a singular term or plural. Other keywords that appeared more than 500 times in documents related to MOOCs were education, teaching, and students. This indicated that MOOCs research was mostly concerned with teaching and learning. Other common keywords appearing more than 100 times were ( motivation. These keywords were clustered mainly around computer sciences concepts. Title analysis. Figure 2 shows the visualization of a term co-occurrence network based on title fields with a minimum of 10 occurrences of a term. We used a binary counting method, wherein the number of times a noun phrase occurred in the title of a publication played no role (van Eck & Waltman, 2014). According to van Eck and Waltman (2014), a noun phrase that occurs only once in the title of a publication is treated in the same way as a noun phrase that occurs, for instance, 10 times. Figure 2 reveals the word "course" was the main term acting as the central node of the whole network (Verk, Golob, & Podnar, 2019)

Major Players and Collaboration in MOOCs Research
This study examined the characteristics of scientific collaborations on MOOCs research by analyzing (a) the countries that most frequently contributed, (b) the main institutions involved in MOOCs research, (c) authorship analysis, and (d) citations analysis.
Countries contributing most to MOOCs research.  Main institutions. Table 9 shows the institutions from which most of the publications on MOOCs originated. Out of the 3,118 documents, Universidad Nacional de Educacion a Distancia (n = 57), which is one of the world's largest universities, located in 13 countries in Europe, America, and Africa, contributed most to publications on MOOCs. This was followed by (    VOSviewer software was used to present a network visualization (see Figure 3) Liyanagunawardena et al. (2013) in the early days of MOOC expansion, as well as another empirical study on how MOOC videos affected student engagement by Guo, Kim, and Rubin (2014). Other documents most often cited were literature reviews or those that addressed the issues of (a) learner disengagement, (b) enrolment and completion, (c) challenges, (d) quality, (e) motivation, and (f) pedagogy.

Discussion
This study was motivated by two observations. First, MOOCs have been regarded as a tool that contributes towards the universal agenda of addressing the digital divide and promoting equity in educational opportunities (Ma & Lee, 2019), as also stipulated in the Sustainable Development Goal 4 (United Nations General Assembly, 2015). Second, MOOCs have also been suggested as a strategy for the internationalization of higher education institutions (Kerr & Reda, 2019). However, questions have been raised as to whether research on MOOCs is interdisciplinary and conducted collaboratively across different parts of the world. Consultation with people from different local contexts and backgrounds represents inclusiveness, which is important in creating MOOCs (King, Pegrum, & Forsey, 2018). To by Liyanagunawardena et al. (2013) and Zancanaro, Todesco, and Ramos (2015). Most research on MOOCs was found in journals and conference proceedings in the form of articles and conference papers.
The documents most often appeared in titles meant for those working in the area of computer science, information systems, or information technology, and based in the U.S. and central Europe. In addition, most MOOCs documents were published in English, despite flourishing MOOCs delivered in different languages (Lambert, 2020). This suggests that the research has paid less attention to MOOCs as encouraging inclusiveness, and has undervalued their important role in promoting part of the United Nation universal agenda, particularly to achieve Sustainable Development Goal 4 in ensuring inclusive and equitable quality education for all by 2030. The underlying implication is that despite the thriving research on MOOCs, it has aimed mainly at a small, focused group. Potential stakeholders from different areas are missing out on the potential of MOOCs, as well as the latest developments, recommendations, and effects.
Regarding the second research question, our observations on subject areas, keywords, and titles suggested that MOOCs research has been confined mainly within the domain of computer science and the social sciences area, particularly as these relate to education. The clustering of MOOCs research in just two main subject areas is further evident by the keywords most frequently used. These indicated that most MOOCs studies were concerned with (a) education, (b) teaching, (c) students, (d) curricula, (e) learning systems, (f) engineering education, (g) online learning, and (h) higher education. This somewhat differs from Ebben and Murphy (2014), who showed that journals publishing MOOCs research lacked penetration into the traditional fields of study such as the humanities, sciences, and social sciences. Perhaps online and distance education journals, and those in computer science, have speedier publication processes due to rapid changes in their subject matter (Ebben & Murphy, 2014).
The narrow disciplinary fragmentation may also due to the complexities of (a) carrying out MOOCs research, (b) framing diverse problems, and (c) aspiring for collaboration (Cairns, Hielscher, & Light, 2020). Therefore, it is important for future research on MOOCs to expand beyond the fields of distance education and computer science into different discipline-based and interdisciplinary research. For instance, a study on employer receptivity to using MOOCs in recruiting, hiring, and professional development (Radford et al., 2014) could create more awareness and use of MOOCs by various bodies or organizations. The expansion of subject matter, key areas, or large-scale field trials in MOOCs research may also help address the problem of abysmal completion rates in MOOCs, and focus on finding what works, where, and for whom (Kizilcec et al., 2020).
As regards the third research question, there seemed to be a reasonable amount of scientific collaboration on MOOCs research across the globe as reflected in our analysis of countries, institutions, authors, and citations. Although MOOCs initially began in Canada, the United States, China, and Spain were the top three countries from which scholarly writings on MOOCs have been published. This finding supports previous studies that found most publications on MOOCs, as well as the vast majority of MOOCs participants, were from North America and Europe (Lambert, 2020;Zancanaro & Domingues, 2018;Zhu et al., 2018a). The U.S. had the highest number of publications on MOOCs, which indicated that it was leading in MOOCs research and, potentially, had directed funding to it. This may be corroborated further by the fact that the U.S. has by far the most top-ranked universities in the world.
The U.S., thus, has been in a much stronger position to bring the best possible content from the best schools and best professors to everyone with online access. In addition, most service providers and platforms for MOOCs originated in the U.S. and Europe, and various initiatives such as European OpenupEd supported the proliferation of MOOCs there. The big gap between MOOCs that originated from these countries compared to the rest of the world should be a point of concern, since one of the Probing the most often cited documents, those from the U.S., China, and Spain were cited most, but our findings also pointed towards a healthy citation impact from different countries around the globe.
Citations of documents from multiple countries implies that authors recognize their scientific community in different geographical areas, which may contribute to forming scientific paradigms (Pan, Kaski, & Fortunato, 2012). Compared to the U.S., China seemed to have had more recognition in terms of citations from central Asia, which coincided with China's aspirations, since the end of the Cold War, to influence this part of the world (Rogers, 2007).
Collaboration is important so that MOOCs may adapt to local contexts (Ichou, 2018)  innovations. Views on the danger of MOOCs in reinforcing inequalities in education (Rohs & Ganz, 2015) and MOOCs that help distribute free education (Lambert, 2020) might also be addressed by forging more collaborations with different stakeholders. This effort will serve to encourage inclusive and equal access to education.

Conclusion
This study used bibliometric analysis to undertake a comprehensive overview format. Findings showed that early research on MOOCs was carried out by scholars from both the East and the West, and has continuously grown and been widely disseminated since then. Nevertheless, most MOOCs research has focused mainly on the same limited fields of computer science and distance education that dealt with topics connected to the social sciences discipline. This has led to disengagement from other disciplines and reduced the emergence of new ideas and innovation. There has been increasing collaboration on MOOCs research among scholars or institutions from a limited group of countries, implying a lack of perspectives from different economic, cultural, and institutional backgrounds. This evolution of MOOCs in general reflects a rising emphasis on open online courses conducted at a global level, thereby addressing the criticism that MOOCs are decreasing. In addition, collaborations and communications involving MOOCs research, which greatly influence educational decisions and perspectives, are confined mainly to certain geographical areas, and do not represent countries in the greatest need for the benefits of MOOCs. A sizeable increase in investment and dedicated funding to encourage stronger international participation from lesser developed nations will be valuable, and is recommended, to ensure that opportunities in MOOCs may be equally enjoyed and appreciated.
It must be acknowledged that this study relied solely on the Scopus database and on the choice of keywords used in document titles. We did not consider other rich databases such as Google Scholar or documents that discussed MOOCs but may have had titles outside our search parameters. Extending the procedures for text analysis to also include abstracts would likely reveal additional information and frequencies. In addition, some authors or institutions might have registered more than one name into Scopus or provided different spellings, and this may have resulted in inaccurate details on authors' affiliations or productivity.
Despite these limitations, this study contributes to a better understanding of the trends in MOOCs research and publications. Each of the indicators points towards growth in this field of research which may offer more opportunities for bettering current educational systems. This study extends and complements previous findings on MOOCs literature by using bibliometric methods. The current analysis produces several exciting observations that clearly highlight the rising importance of MOOCs in the educational environment around the world, as well as their dissemination, and the need for more research involving cooperation among various regions and different fields. More studies are needed to explore and help balance the education gap that may arise in the context of MOOCs development.
Focusing efforts on cultural differences in MOOCs is one likely topic to be pursued. This will facilitate the attainment of educational goals worldwide and ensure that everyone may benefit from MOOCs.