Social Networks and the Building of Learning Communities: An Experimental Study of a Social MOOC

This study aimed to analyze the student’s behaviour in relation to their degree of commitment, participation, and contribution in a MOOC based on a social learning approach. Interaction data was collected on the learning platform and in social networks, both of which were used in the third edition of a social MOOC course. This data was then studied via statistical methods and analysis of social networks. This study assumes that social communities would arise around the course, would remain over time, and that participants would even contribute with new proposals. The findings indicated that social learning communities are built and continue only while the course is open and while the teachers are involved in fostering participation. Although this study is limited, the design criteria of the course, the pedagogical model on which this is supported, and the methods applied for this analysis provide other researchers and educators with clues for better understand the dynamic process of social learning in social MOOCs.

Open and distance learning seeks to make education more open to those who need or wish for alternative opportunities to the traditional system (D'Antoni, 2009). As well, "it also draws upon open technologies that facilitate collaborative, flexible learning and the open sharing of teaching practices that empower educators to benefit from the best ideas of their colleagues. It may also grow to include new approaches to assessment, accreditation and collaborative learning" as claimed the Cape Town Open Education Declaration (2007, para. 4). Brahimi and Sarirete (2015) note that in the traditional learning model, the teacher serves as the repository and transmitter of knowledge. In the MOOC's approach, however, the teaching model is generally inverted and focused on the learner who chooses his or her own "roadmap" by interacting with peers and accessing, in a flexible way, all information and resources (open and free). The learning cycle is an ongoing process that must be designed to improve the quality of knowledge as well as the collaboration among learners (Brahimi & Sarirete, 2015).
MOOCs are likely the leading exponent of the new Education 3.0 model. Education 3.0 is an umbrella term used by educational theorists to describe a variety of ways to integrate technology into learning. Although there is not a consensus definition for this term, most educators agree that Education 3.0 is a connectivist, heutagogical approach 1 to teaching and learning where teachers, learners, networks, connections, media, resources and tools create a unique entity that has the potential to meet individual learners', educators' and even societal needs, being that most resources are free (Gerstein, 2013). Education 3.0 suggests a new pedagogy where educators and learners create, shape, and evolve knowledge together, deepening their skills and understanding as they go (Cape Town Open Education Declaration, 2007).
Internet and technology, in general, and social media, in particular, have changed learning processes by questioning not just where people learn, but also how and why they learn. Education 3.0, and MOOCs as well, propose a relevant change in teachers' and learners' roles. Learners play a key role as creators and connectors of knowledge that are shared with others through social networks. Learners become the authors, drivers and assessors of their experiences and educators become a guide-as-the-side, coach, resourcesuggester and cheerleader who encourage learners to create their own learning journey (Gerstein, 2014) and knowledge.
At the same time, social networks have created conditions for the development of new paradigms and methodologies in education (Putnik et al., 2015) such as community-based learning or social learning.
Social learning is a key aspect of MOOCs platforms which supports scalable peer-based learning as well as acting as the main channel of interaction between teachers and students (Brinton et al., 2014) and between learners themselves. Sol, Beers, and Wals (2013) define social learning as "an interactive and dynamic process in a multi-actor setting where knowledge is exchanged and where actors learn by interaction and co-create new knowledge in on-going interaction" (p. 37). On the other hand, Paredes-Labra, Herrán-Gascón, and Velázquez-Vázquez 1 a holistic approach to developing learner capabilities, with learning as an active and proactive process, and learners serving as the major agent in their own learning, which occurs as a result of personal experiences.

42
(2012) stated that learning communities (and communities of practices as well), by exchange of ideas, experiences, proposals, and so on, must promote the improvement of professional and academic practices and the development of learners and professionals in a collaborative and integral way. Thus, active participation, socialization of experiences and the approach to common issues of interest are all key aspects to the success of this tool. In addition, Putnik et al. (2015) stated that social networking concerning education issues must be applied in two different dimensions: (1) to organize the educational process as a social network-based process (useful for peer assessment, discussions, and collaborative tasks); and (2) to analyze the students' interactions using techniques from the social network analysis field (SNA).
SNA is an important part of the social networks theory and corresponds to the study of relationships of individuals or groups of individuals. SNA represents both actors and relationships in terms of network theory, depicting them as a graph or network, where each node corresponds to an individual actor within the network, e.g., a person, a group or an organization, and each link symbolizes some form of social interaction between two of those actors, e.g., friendship or kinship. This representation is called sociogram.
SNA draws on various concepts from the graph theory to describe and illustrate the individual and collective structure of a network. The most relevant are the centrality measures (Freeman, 1977) which allow us to identify the most prominent actors: that is, those extensively involved in relationships with other network members.
An advantage that sociograms present is that they are highly meaningful and easily usable since they generally allow users to rank and filter network nodes or links based on the values of network concepts. Even more, sociograms are able to represent different features by designing elements using different shapes and colours, which enable teachers to gain insights in a glimpse (Zhu, Watts, & Chen, 2010).
SNA techniques have been applied to data from very different contexts, which allow researchers to model different types of interactions. In the educational arena, Dawson, Tan, and McWilliam (2011) used centrality measures to monitor the learners' creative capacity; Putnik et al. (2015) studied the correlation between these measures and students' performance; and Rabbany, Elatia, Takaffoli, and Zaïane (2014) applied SNA so as to analyze the structure of the interactions raised between the students in forums.
Social networks and collaborative learning are essential elements of a MOOC, especially for massive courses called "cMOOC." The use of both internal and external social networks in a MOOC is well appreciated by students (Castaño-Garrido, Maiz-Olazabalaga, & Garay-Ruiz, 2015), and their use is also very suitable for developing collaborative learning tasks (Brinton et al., 2014). On the other hand, sociograms and centrality measures are valuable tools for understanding the social network phenomena (Zhu, Watts, & Chen, 2010). Therefore, these tools along with analytical techniques are used to analyze the students' behaviour in relation to their degree of commitment, participation, and contribution in a MOOC supported on a social learning approach.

MOOC's Pedagogical Approaches and ECO Project
To introduce connectivism as an essential element of a learning model is to assume that the learning process is no longer an internal or individualistic activity (Siemens, 2014). ICTs, Internet, and other digital tools are making an impact on the field of education and are forcing the education community to develop new learning approaches based on interaction, exchange, and collaboration, and focused on learners rather than on content or on teachers as has been done in the past. In this sense, connectivism supposedly has the potential to improve teaching methods and encourage both innovation and new pedagogical practices (Yuan & Powell, 2013). Siemens (2012) used the terms cMOOC and xMOOC to contrast two different pedagogical approaches of MOOCs. While cMOOCs are rooted in principles of connectivist learning that emphasize creation, creativity, autonomy and social networked learning, xMOOCs are instead based on a transmission model of teaching and learning. In addition, Yuan and Powell (2013)  ECO Project is aimed to design, develop, and evaluate different MOOCs for educators and the teaching community in order to provide basic tools for both the development of these types of courses and to promote specialization in different areas. Furthermore, ECO Project aims at creating courses that follow a new pedagogical approach.
The pedagogical model of ECO Project assumes a "traditional" educational design in combination with the power of social media, bringing together didactic content, participation, debate, and dialogue through internal (forum) and external (Facebook and other) social networks. In any case, this pedagogical approach, based on connectivism and the social-constructivist perspective, goes towards the construction of a learning community that could be active even after the course has ended. So, the "s" in ECO sMOOC stands for "social," since it should (ideally) provide a learning experience marked by social interactions and participation among students (Morgado et al., 2014).
According to Wenger (1998), communities of practice are characterized by three essential dimensions that make them different from other kinds of groups or communities: what it is about, how it functions, and what capability it produces. In this sense, "communities of practice develop around things that matter to people. As results, their practices reflect the members' own understanding of what is important . . . even when a community's actions conform an external mandate" (Wenger, 1998, p. 4). Furthermore, and 44 independent of the nature of the community, its development depends on a diverse and well-distributed internal leadership (Wenger, 1998).
Based on that, in ECO's courses, the use of internal and external social spaces aims to go beyond the simple interaction between people (chat) and delve into the thoughtful exchange of good-practices, thus creating a network of interest, knowledge, and new ideas. In order to reach this goal, and following ECO's pedagogy, some collaborative (non-mandatory) tasks must be regularly proposed using social networks (e.g., forum and Facebook). The main objective here is to generate not only discussion, but also further debate between participants around each topic.
Because these courses can have a wide variety of targets, it is not possible to design a one-solution-fits-all model. It needs to be designed as a framework model within which choices are made and specified in order to make the courses effective. People who participate in this learning event become part of a learning community, which, to some extent, is also a community of interests or a community of practice (Morgado et al., 2014).
In short, the main pedagogical principles of ECO Project model are (a) learner-centeredness, (b) autonomy, (c) interaction (as "s" in sMOOC indicates), (d) flexibility, (e) digital inclusion and (f) ubiquitous learning.
Thus, the xMOOC's pedagogical approach is avoided by applying connectivist and social-constructive learning and situated practices.
This new pedagogical framework highlights a new scenario in which social networks have a special importance. Thus, analytical studies addressed to evaluate whether the goals of this learning model are achieved or not must be carried out.

Methodology
This paper adopts a case study approach to investigate our research questions. This method of study is especially useful to test theoretical models by using them in real world situations. While it will not answer a question completely, it will give some clues and allow further conclusions to be drawn on the subject. As Yin (2009) claimed, the case study research is a linear but iterative process, meaning that each step in the process must be reviewed and each decision questioned. This iterative analysis has been performed for this research at the beginning of each edition of the course. This study will explore the following research questions related to social network-based learning and the building of learning communities: RQ1. What is the profile of the learners who follow a social MOOC and what is their behaviour?
RQ2. What types of networks emerge?
RQ3. Do social network-based activities enable the creation of learning communities that endure over time?

Data Sources
Our study comprises different data sources. On the one hand, MOOC participants' activity data collected in ECO's platform was read and stored in a MySQL database. This data provides us with information about videos watched entirely, completed peer-review assignments, and number of messages written and replied to by each participant in the forum. On the other hand, activity registered in Facebook was extracted using Facebook Graph API and saved in the same relational database. From Facebook, we recovered all the posts written in response to each topic and its social valuation ("likes"). The account created on Facebook is public and is called "Innovacion Educativa y Desarrollo Profesional." Due to ECO Project's awareness of legal and ethical issues concerning the use of personal data and its commitment to reduce technological barriers for those users with special needs or at risk of exclusion, the platform does not oblige users to fill out any feature except for a name. Therefore, two anonymous surveys were prepared: an initial one, focused on discovering the learner's profile, their background, and their goals in this course; and a satisfaction survey at the end of the course to measure to what extent the course fulfilled its aims.

Analytical Tools
This study employed two analytical techniques: SQL/OLAP and SNA (Social Network Analysis).

SQL/OLAP, to manage and aggregate data from different points of view and pivot tables (available in
Microsoft Excel) to visualize data and draw meaningful plots. ORA Software (Carley, 2014) was chosen to measure and visualize the social networks, which arose from the participants' interaction in forums and on Facebook. Specifically, centrality measures at an individual level were used to identify the most influential participants in the network.

Research Context: Course Description
The course being analyzed is called "Educational innovation and teacher professional development.
Possibilities and limits of ICT." It was taught in Spanish, hosted in OpenMOOC platform (https://ecolearning.eu/), and lasted eight weeks. The course was aimed at preparing students to design, develop, and evaluate contextualized educational innovation projects, which included the use of information and communications technology (ICT).
The course was designed under a social learning approach and seeks to encourage students to work autonomously (i.e., to be able to independently select activities and content that are interesting for them) as well as to stimulate the interaction between participants through peer-review tasks and debates opened in the forum (inside the platform) and social networks (Facebook, Twitter, Google+/Hangouts, in this case).
Peer-review activities were compulsory (and required to obtain badges), whereas discussions were optional.
Furthermore, at the beginning of each content module the teaching team invited participants to watch a Hangouts meeting in order to present the team, structure, content, activities, and assessments of each unit.  P2P tasks include individual effort, drafting of a report, and collaborative effort via a peer-to-peer assessment task After reviewing basic and supplementary materials (articles and videos), participants should debate unit content via ECO´s platform forum Exchange of experience on Facebook related to students' participation in educative innovation projects Unit 5 -weeks 6 and 7 P2P tasks include individual effort, drafting of a report, and collaborative effort via a peer-to-peer assessment task After reviewing basic and supplementary materials (articles and videos), participants should debate unit content via ECO´s platform forum Exchange of information on Facebook related to professional educative innovation networks Unit 6 -week 8 Satisfaction Survey

RQ1. What is the Profile of the Learners Who Follow a Social MOOC and what is their Behaviour?
Firstly, we analyzed the initial survey using Google Forms (see an excerpt of this survey in Figure 1). We   In general, learners who started the first few weeks are those who followed and participated in the course, although as can be observed in Figure 3, there were also participants who enrolled later and performed the compulsory activities. On the other hand, the fact that these kinds of courses are both free and flexible also influences the degree of commitment, which is generally lower compared to a traditional setting. As commonly occurs (Putnik et al., 2015), those who participate more achieve a higher mark, as shown in  Regarding the activities performed in the forum and on Facebook, data shows that the highest number of messages were written during the week in which the discussions were opened, though the discussions continued until the end of the course, with the exception of activity number 5 in which most comments appeared two weeks later, most likely due to the delivery of the second peer-reviewed task. Figure 5 displays the number of messages written in each forum activity each week except forum activity 1, which was discarded because it was focused on introductions and not on discussion topic. Figure 6 depicts the number of comments written on the wall. These plots show that both tools were used equally and that the activities were implemented at the same time. Thus, it can be claimed that the tools were not a barrier for the learners.

52
Finally, another interesting fact (see Table 2) is that participants contributed mostly on weekdays, considering that the number of messages was remarkably lower on weekends.

RQ2. What Types of Networks Emerge?
Although papers such as Alario-Hoyos et al. (2014) demonstrate that learners in cMOOCs prefer forums, this study conversely does not, as shown in Figure 5 and 6. Social MOOCs attract people who like and frequently use social network sites and thus, the activity in this respect is quite balanced. The total number of participants, both learners and teachers, was 104 in the forum and 103 on Facebook, and as can be observed in Table 3, their participation per weeks is quite similar. One limitation of this study is that we could not identify each person on Facebook with his/her account in ECO Platform thus we could not know for certain that they were the same person.. Only those who filled the ECO register form properly and were identified by their full name on Facebook were able to be followed (35 people). These students participated in both platforms but in different activities. Thus, the tool was not a barrier. Regarding the number of messages, Table 4 reflects that the participation is quite similar. In fact, most learners only wrote one message per discussion task and they generally did not contribute in all of them.
Furthermore, it must be taken into account that some participants instead of writing a long text, sometimes sent several short and consecutive messages, which leads to this small difference. The average length of messages is 400 characters, ranging from 44 up to 3110 characters. Thus, these texts are quite extent and discursive, although regrettably they do not reach the metacognition phase which is necessary to construct new knowledge according to Kellogg, Booth, and Oliver (2014). Basically, learners paraphrase the content of a video or text. For instance, Por lo general, por lo que comentáis, la falta de disponibilidad de equipo y recursos no es la barrera más importante para la incorporación de la innovación con TIC;. . .
You usually say that the lack of availability of equipment and resources is not the most important barrier for including ICTs in innovation. . . Sociograms, in Figures 7 and 8, depict the interaction produced in each Facebook and Forum task respectively, and show the participants who wrote comments in each discussion. Red nodes represent learners and teachers and yellow nodes, topics. The links between nodes signify that a comment was written by that participant regarding that topic. There were 27 people who wrote messages in the five posts on Facebook, 9 people in four, 7 in three, 20 in two, and finally 35 in only one post. Regarding the forum, a similar type of network appears. In this case, 28 people contributed in the four forum activities but only once, 14 in three, 15 in two, and 30 in only one forum task. The thickness of the edges in Figure 8 depicts the number of messages sent. This shows that only three people wrote several messages (@jgdq314242, @Elia, and @Carlos.Rodriguez) but, as can be observed, only in one or two activities. It must be highlighted that all of them were teachers.
Tools that build these sociograms allow teachers to filter by different features such as topics or number of messages interchanged, enabling them to discover who is interested in each topic (community of practice) and to what extent.  If we focus now on the degree and type of interaction, Figures 9 and 10 depict that the social network created is radial, which means that teachers were the ones who encouraged and kept the activity going. Innovacion Educativa was the account used by teachers on Facebook. The thickness of the edges in the sociogram 55 represents the number of messages interchanged. It ranges from 1 to 5. It can be observed that the interchange of messages in the Facebook community was very low, only pointing out the learner Javarteacher Arias Bon who sent several messages to Concha Torres. Furthermore, few participants communicated directly with others: this is the case of David Armas Young, Noelia UCantabria, and Irene Echevarría.
The forum community ( Figure 10) seems to have facilitated the communication between more people. That is to say that participants in forums answered comments from other students as well as those directly from teachers, whereas participants on Facebook generally responded only to the main comment. Despite this, the network is also radial. Learners @joselbd, @inakidg319341, @Ireneeg322447, and @javarbon322230 communicated with more participants, as the number of links that emerge from their nodes show.  Because this is more difficult to appreciate graphically, SNA centrality measures were used (Freeman, 1977), allowing us to know which participants contributed the most and which were essential to keeping the network alive. Figures, thus, help to measure, to rank, and to compare the differences in participation.
The degree measures the immediate influence: which participants were more active, i.e., those who were more deeply involved in the activity, either asking or responding. If we take into account the direction of the links, the indegree reveals those participants that responded the most, because of their reputation or due to the question itself, i.e., it may be an interesting or controversial question. On the other hand, the outdegree depicts those participants who respond the most, i.e., those who contribute to the activity by adding new content, proposing new approaches to the discussion, or asking for further explanations. The betweenness detects those participants who discuss the most and make others discuss, either by asking questions or answering in such a way that the others are encouraged to intervene, i.e., they act as the vehicle of communication. Authorities are those participants with high authority values who seem to have some sort of reputation, because what they ask is usually worthy of a response. Tables 5 and 6 show the SNA measures of the people whose participation was remarkable (marked in bold) on Facebook and in the Forum respectively. As can be observed, the leadership on Facebook was held by teachers (mainly by Innovacion Educativa y Desarrollo Profesional account), since they had the highest values in indegree, outdegree, and authority. It must be also remarked that the learner Javarteacher Arias Bon had facilitated and encouraged participation in the learning community as his betweenness and indegree show. This leads us to confirm that the Facebook learning network is kept thanks to teachers' involvement. Regarding the forum network (Table 6), it was the teachers who received most of the comments, whereas learners (e.g., @joselbd, @Ireneeg322447, @bea_147320965, and @javarbon322230 with high outdegree) were the ones who answered the most. The teacher, @jgilq314242, played an important role as a link between participants as his betweenness indicates, and @jgilq314242 was an active person whose comments deserved an answer.

58
On the other hand, Facebook also provides us with social valuation about learners' contribution by mean of "like counts." Figure 11 highlights participants whose comments were considered as relevant to their peers.
This figure also depicts that a reputation system helps to build subnetworks around the people with certain social valuation such as Javarteacher Aries Bon and Cristina Sánchez. Figure 11. Interaction graph built from the interchange of messages on Facebook. Node size is proportional to the number of likes that these participants received. This number is shown in Table 7.

RQ3. Do Social Network Based Activities Enable the Creation of Learning Communities that Endure Over Time?
One of the main goals of ECO Project is to increase the number of well-qualified teachers in Europe and to enable them to create their own online courses and other open educational resources to be distributed through its open learning platform. Thus, its aim is to build a learning community around "educational issues." To achieve this goal, ECO courses are always open and encourage students from one edition to participate in the following ones. From this participation, it is assumed that new groups of people with similar interests will appear and that those groups will provide new courses designed under this pedagogical model.
Although the experience is very short, only one year, our experience has not been very positive. None of our learners from one edition participated in the following ones. Basically, students followed the course as far as they could or were interested. No new course or proposal arose from this group of participants. Thus, the participants of our communities of practice, according to Wenger (2010), only achieved the engagement level and not the alignment level needed to negotiate and plan a new course.
In analyzing the satisfactory survey (46 completed), which tried to assess the degree of conformity of the course to the pedagogical model and degree of satisfaction of the learners, we can say that 89% consider that the course boosts interaction between peers and feel very satisfied with the following issues: (a) support from other students; (b) posts and comments made by other students, (c) projects, jobs, or other resources shared by other participants, and (d) the feedback and comments others made about their work. This drives us to think that the model is suitable for, at least, people with a profile similar to the one described in this paper. Regarding the reason why they enrolled and followed the course, learning new things was the most selected option (77%), along with achieving a certificate (50%).

Discussion
ECO Project, based on social learning, aims (among other issues) at creating a learning community for educators and teaching and encouraging learners to contribute their knowledge, experience, and resources among others. Likewise, ECO Project offers participants the opportunity to become promoters of new courses following this new pedagogical approach. In fewer words, ECO's principles are aimed to promote and boost the change that Education 3.0 suggests.
After the completion of the analysis of the third edition of "Educational innovation and teacher professional development. Possibilities and limits of ICT" course, we found out that participants have not contributed to building the autonomous and sustainable learning community that ECO Project proposes. However, the study's analysis, although limited, allows us to draw several conclusions.
Firstly, the removal of barriers that facilitates the enrolment of people (simply with a mail account) prevents teachers from carrying out an in-depth-analysis of these courses, and thus, the process of continuous instructional improvement is restricted. On the other hand, this simplified registration system used by ECO Project, and the fact that the course is always open, favour the incorporation of new learners throughout the course, which leads to growing the learning community with other points of view and expertise.
Nevertheless, in our analysis, we found that the students' profiles differ from other MOOCs (Alario-Hoyos et al., 2014) in two aspects: (a) they use the forum and Facebook indistinctly and are thus socially-connected and (b) the degree of participation is quite reduced and teacher-centred. Regrettably, their learning model is still traditional, meaning that they continue bringing their knowledge to those who know more with the aim of being given feedback, as Gerstein (2014)  Another reason could be the short duration of the course, which could limit the creation of strong links. The study carried out by Kamalodeen and Jameson-Charles (2016) shows that social ties started to be centred on one participant after eight weeks. They also mention that their learners preferred viewing to contributing, and thus a community and collaborative spirit must still be developed to achieve the goals of Education 3.0.
In addition, we must highlight that most participants (62.1%) had not participated in another MOOC before.
This fact leads us to think that learners' lack of experience in courses of this sort (and likely also in communities of practice) has hindered the autonomous and independent interaction among participants.
In this sense, we can remark as a limitation of the course the absence of strategies in order to promote a culture around communities of practice: concept, importance, transcendence, advantages, functions, developing, etc. The belief that this community might arise "spontaneously" led the teacher team to take over the leadership of all debates during the course. This fact is contrary to Wenger's (1998) proposal about the need to recognize people/leaders that could legitimize the community as a place for sharing and creating knowledge, working from the inside rather than attempting to manage communities from the outside.
It is interesting to note that the majority of the participants that did not complete mandatory tasks were the ones who participated more actively on social spaces. Even more, some discussions were kept active throughout the course and drove participants to reflect and re-think issues, one of the landmarks in learning. This demonstrates that people who enrol in social MOOCs do not seek to achieve a certificate but to gain knowledge on a topic, as the results of our internal satisfaction survey confirm.
Likewise, we would like to highlight the usefulness of sociograms for analyzing the kind of communities and sub-communities built, whether the activity is centred on the teacher or not. Sociograms also allow teachers to know what topics attracted more attention and, if observed in different snapshots, how the networks evolved. On the other hand, SNA measures enable teachers to identify prominent people in the social communities created. This information could be used for different purposes such as the creation of work teams for the purpose of joining more and less-connected people, or to assign new issues to more active people for their dissemination and discussion. Some SNA metrics such as betweenness or authority could also be used as elements of gamification since they reveal the importance of these actors in the network.
We found that social acknowledgment is a relevant aspect of social interactions, especially in social networking websites. For this reason, social networks in general have a Karma System that points out (positively or negatively) users' reputation. Moving to educative aims, a Karma System can be useful to indicate "participants' level and quality of course engagement" as well as to "encourage interactivity and high quality submissions" (Morgado et al., 2014, p.29) from participants, and consequently enable the apparition of learners' communities around influential students. In this sense, we found that Facebook's reputation system helped to build subnetworks around the people with a certain social valuation. But, in any case, those subnetworks have generated an autonomous learning community as ECO Project's objectives deal with. Even more, no learner from previous editions of the course participated in the following ones. Unfortunately, we have not managed to generate a strong learning community either during the course or at its completion: the networks were created around teachers' feedback, learners basically commented once per topic and, after the course ended, people did not return to Facebook or to the forum to participate. As Wenger (2010, p. 182) claims, "the engagement in practice is rarely effective without some degree of alignment with the context" and this must be the main cause of our failure. Most learners were interested in learning but were not motivated enough to become drivers of a new MOOC course or to build a learning community on their own.