PAPER A FRAMEWORK FOR COLLABORATIVE NETWORKED LEARNING IN HIGHER EDUCATION: DESIGN & ANALYSIS A Framework for Collaborative Networked Learning in Higher Education: Design & Analysis http://dx.doi.org/10.3991/ijet.v9i8.3903 Ghassan F. Issa, Haya A. El-Ghalayini, Ahmad F. Shubita, Mohammed H. Abu-Arqoub University of Petra, Amman, Jordan Abstract—This paper presents a comprehensive framework for building collaborative learning networks within higher educational institutions. This framework focuses on systems design and implementation issues in addition to a complete set of evaluation, and analysis tools. The objective of this project is to improve the standards of higher education in Jordan through the implementation of transparent, collaborative, innovative, and modern quality educational programs. The framework highlights the major steps required to plan, design, and implement collaborative learning systems. Several issues are discussed such as unification of courses and program of studies, using appropriate learning management system, software design development using Agile methodology, infrastructure design, access issues, proprietary data storage, and social network analysis (SNA) techniques. Index Terms—collaborative learning; networked learning; social network analysis; network design; learning community. I. INTRODUCTION This paper is based on a proposal for the implementation of a major project that connects several higher educational institutions in Jordan in a collaborative networked environment. The paper will formulate the implementation steps into a general framework that can be easily adopted by similar projects. Jordan is considered one of the leading countries in the Middle East in higher education, and the first in the area to implement privatization of quality higher education. However, in the past few years, several factors have been negatively affecting higher education in Jordan including the incompatibility of learned skills and those required by employers, outdated teaching and learning methods, and the inefficient use of technology in education [1]. Development of Jordan society in recent years is going through major transformations in terms of socio-economic and political issues, change of priorities in life, moral concepts, and moral norms. Social changes posed new challenges for higher education requiring a revision of traditional teaching methods, technologies, and services. The ongoing reform of the educational system in Jordan aims at creating the necessary conditions for achieving high quality learners, who will successfully interact in a national and international cooperative environment. However, we believe that reform efforts must take into account some serious challenges related to learning, research, community service, and innovation. Currently, higher 32 educational institutions in Jordan suffer from the lack of collaboration or linkage with secondary educational system, industry, research groups and research centers, and innovations. There are hardly any existing learning community spaces that engage policy makers, educators, civil societies, or the new generation of millennium students. Needless to say, there are virtually no existing learning communities, nor a reasonable technological infrastructure that facilitate proper collaboration between existing higher educational institutions. To overcome the above mentioned shortcomings, this paper presents a framework that takes advantage of the latest development of Information and communication technologies (ICT), and social networking to connect several higher educational institutions in Jordan in a collaborative networked environment. The implementation of the proposed framework aims at enhancing higher education from different perspectives including policy makers, administrators, teachers, researchers, and students. The proposed framework employs ICT to enable educators to adopt new learning methods for the realization of modern concepts in education based on learner-centered approaches utilizing social constructivism learning theory such as: network-based learning, problem-based learning, collaborative learning, and competition-based learning [2], that reflect the basic ideas and principles of humanistic pedagogy. The work introduced in this paper is based on Vygotsky’s theory [3] which states that collaborative learning is a cognitive development that depends on full social interaction which leads onto online learning communities that include interactivity, social context, and technologies. Additionally, this research relies on the community of inquiry framework [4, 5] for defining blended learning. In the next section we will shed some light on some of the major issues that relate directly to our proposed project including blended learning, collaborative learning, and networked learning. II. BACKGROUND A. Blended Learning Blended learning or as sometimes referred to as "hybrid", "technology-mediated instruction", "webenhanced instruction", or "mixed-mode instruction", is a student centered (constructivist) approach based on faceto-face and computer mediated learning which blends collaborative and self-reflective learning activities. Blended learning employs traditional methods such as http://www.i-jet.org PAPER A FRAMEWORK FOR COLLABORATIVE NETWORKED LEARNING IN HIGHER EDUCATION: DESIGN & ANALYSIS lecture and group-based learning with new approaches of online and networked learning through social media, thus observing individual differences in learning methods and knowledge accumulation. Hope [6] identifies a set of factors for the prevalence of blended education such as: motivation of social communication among students, students and teachers, and teachers themselves; enhancement of learning achievements of students, by taking into account their pace, capacities and circumstances at any time and place; enhancement of students' ability to work in small groups and in a team spirit; enhancement of the quality of teaching materials, and teaching and learning skills. Osguthorpe & Graham [7] on the other hand, suggest that blended course structure varies depending on the goals that are included in the design, where the goals can be one or more of the following: pedagogical richness, access to knowledge, social interaction, personal agency, cost effectiveness, and ease of revision. Community of Inquiry framework views blended learning from a socioconstructivist perspective [4, 5]. In order to create a learning community, this model defines three major elements: teaching, cognitive, and social presence. According to this model teaching presence provides the structure to a course which directs cognitive and social presence. Alternatively, the Blended Learning Curriculum (BLC) model presented by Huang, Ma, & Zhang [8], defines three characteristics of blended learning as follows: flexible learning resources, learning diversity, and e-learning experiences. BLC model focuses on personalized learning and just-in-time transfer of skills. B. Collaborative Learning Collaborative learning by its nature depends on the sociocultural and activity theories [2, 9]. There have been several attempts by researchers to define collaborative learning from different perspectives. Some researcher defined it based on the skills that is shared and transferred from one learner to an other in a group setting [10], others proposed a definition based on knowledge creation and sharing among learners [11]. There is also a definition that focuses on the methodologies and environments that facilitate sharing of common tasks [12]. One last definition that is worth mentioning states that collaboration learning is a "coordinated, synchronous activity that is the result of a continued attempt to construct and maintain a shared conception of a problem”. This definition assumes that learning is greatly enhanced when the learners exist in a collaborative environment [13]. All of these definitions agree with the fact that people can learn through rich social interaction, as mentioned by [14]. A by-product of research on collaborative learning is the concept of learning community (LC) which consists mainly of three major components: interactivity, social context, and technologies [15]. Accordingly, collaborative learning can be considered as interaction between a learning community members for the purpose of learning [16]. Learning communities is said to play a major role in the success and persistence of higher education [17]. A separate field of research has emerged as a new educational paradigm, where computer-supported collaborative learning (CSCL) plays a significant role in enhancing the learning process [18]. CSCL research is focused on iJET ‒ Volume 9, Issue 8: "Learning in Networks", 2014 how learners interact together using computer mediated communication techniques. CSCL provides better understanding and more effective techniques to the applications of blended learning [19]. C. Networked Learning Networked learning is a new form of CSCL that uses what is called "Network Environment" (connections), which assists collaboration between a group of learners; instructors and learners; educators; a learning community and its resources. Therefore, the network environment can help the participants in extending and developing their capabilities and understanding in ways that are meet their aspirations [20]. Networked learning is related to theories of distributed cognition [21, 22] and is rising with the emergence of Web 2.0 technologies. Thus a learning environment is basically a physical environment that facilitates social interaction which allows learning to be distributed over space and time. From this concept, the central concern of networked learning is the argument about the relationship between using a technology and designing that technology [23]. The concept of networked learning has been described as "social intelligence design" based on three different aspects: mind, society, and matter [24]. Some researchers believe that the interaction and collaboration through a network transforms learning, memory, and intelligence from the individual level to the social network level [25, 26]. III. BUILDING A COLLABORATIVE NETWORKED LEARNING FRAMEWORK Building a collaborative networked learning system for higher education is quite challenging. The process requires real collaboration and cooperation between involved universities. In order to proceed in such a system properly, software engineering methodology, namely, AGILE methodology is applied. The advantage of using Agile rather than the more traditional Water fall approach is that the former relies more on customer collaboration rather than contract negotiation. This concept seems very reasonable for our project which involves experts in education, administration, and information technology. In summary, the resulting system reflects the collaboration and cooperation efforts by those experts representing different higher educational institutions. The system's requirements and objectives can be stated as follows: Design and implement a comprehensive networked learning environment between several universities located in Jordan. The system must be able to create a learning community capable of collaborating at all levels: administration level, program level, instructor level, course level, research level, and student level. Figure 1 illustrates a general framework showing the interaction between major system’s components in a networked learning community. The system must be designed bearing in mind proprietary data to be located at each university servers. New contents generated as a result of collaboration such as unified courses are to be located in the central server, which is shared by all participants. The system must provide users with a full range of collaboration, communication, and learning tools. Security issues are important including user authentication, role assignments, and access privileges. The system must be flexible, upgradable, interoperable, and must accommodate new users, and new institutions that will be added in the future. 33 PAPER A FRAMEWORK FOR COLLABORATIVE NETWORKED LEARNING IN HIGHER EDUCATION: DESIGN & ANALYSIS Network analysis tools must be designed and made available to users and to administrators. Social networks are used to examine how institutions interact with each other, characterizing the many informal connections that link instructors together, as well as associations and connections between individual students at different institutions. In addition to SNA other methods such as online surveys and questionnaires are to be designed to measure the satisfaction of users at all levels. We can summarize the objectives of the proposed system as follows: 1. Construct a higher educational collaboration network where decision makers, teachers, students, and researchers constitute a learning community. 2. Enhance learning pedagogy for undergraduate and postgraduate courses. 3. Share existing study programs, courses, resources, and activities. 4. Share and develop partnerships in research and project development. 5. Provide an environment for socialization. 6. Share news, events, and announcements. 7. Adopt best practices in teaching and learning according to national and international standards. The next sections will introduce the design methodology using Agile approach. IV. The following, shown in Table 1, are the preliminary project milestones and deliverables established for our project: TG1: Epics are defined including architecture at high level. Epics are a very large user scenario of the project that is eventually broken down into smaller scenarios. TG2: Overall scope of the project is defined and estimated and our team can start detailing the project plans. B. Using Moodle 2.x LMS Moodle is a free open source learning management system written in PHP, and is deployable on most operating systems including Unix, Linux, Windows, and others. Moodle design is based on the constructivist and social constructivist approaches of learning thus providing a flexible environment for learning communities. Developers using Moodle have full access to the source code, Moodle Documents, and Help. There are several aspects of Moodle that makes suitable for this project. Moodle has an easy, responsive, and personalized user interface suitable for desktop and mobile applications. A comprehensive set of collaboration tools and activities such as forums, wikis, glossaries, and database activities. DESIGN STAGE Given the problem definition in the previous section along with the specified objectives, this section will focus on the design process using the Agile software engineering development methodology. This Agile methodology seems very suitable for our project as it relies more on customer collaboration rather than contract negotiation. A. Agile Software Development Methodology The Agile movement proposes alternatives to traditional project management and software development methodologies. Agile approaches are typically used in software development to help businesses respond to unpredictability [27]. One of the most important differences between the agile and sequential approaches is that sequential features distinct phases with milestones and deliverables at each phase, while agile methods have iterations rather than phases [28]. Agile development methodology helps in assisting the track of a project during the development lifecycle. This is done through regular pieces of work called iterations, which must present a balanced product increment [27]. Agile project plans are based on features. Figure 2 shows the Agile software development milestones. Agile plans projects when main features will be submitted to construction and the most recent iteration tends to have additional details [29]. Agile project plans are arranged into time-bound iterations from 2 - 4 weeks in length [28]. All of the compulsory work, from idea generating to a fully working product, is completed without preventing the work from being done in parallel. This gives stakeholders a better idea of the project progress, as they can use the end result when it becomes available [28]. 34 Figure 1. Framework for Networked Learning Community Figure 2. Agile Software Development Milestones [29] TABLE I. PRELIMINARY PROJECT MILESTONES WITH DESCRIPTION Budget approved for Speculation phase Milestone Resource is assigned for Initiation phase to be completed Initial product backlog complete TG1 Epics (user scenarios) are defined including architecture at high level Speculate phase complete TG2 Overall scope is defined and estimated, Resource available, team can start detailing the plans http://www.i-jet.org PAPER A FRAMEWORK FOR COLLABORATIVE NETWORKED LEARNING IN HIGHER EDUCATION: DESIGN & ANALYSIS An advanced calendar management system which can be automatically populated with new events, due dates and deadlines and which works on the system's level, user's level, or course level. Advanced and simple to use features for administrators that include: customable site design and layout, detailed role and permission management, support open standards, bulk course creation and easy backup, detailed reporting and logs, high interoperability, and security. Flexible course development features including: choosing appropriate pedagogy (instructor-lead, selfpaced, blended, or online), course collaboration tools, embedding external files, group management, multimedia files, marking and grading flexibility and in-line marking, peer and self assessment, and outcomes and rubrics. Ability to observe peer's activities using online users block, and recent activities block. Connection between participants using email, online messages, chatting, and push notification. One of the major advantages of Moodle 2.x is the ability to connect several sites using Moodle as one network, thus creating what is called Moodle Community Hub (Fig. 3). Users can access all the nodes in the network transparently but securely (based on their privileges) in a single sign on. C. Unified Course Approach One of the objectives for the collaboration between different institutions is to learn and use best practices on all levels of learning. Involved institutions sit together and share their experience starting from the program of study level, going through course contents, and ending with course material. The result of this type of collaboration should be transformed into what we call a unified course. While the term “unified” could be misleading, we can define it as follows: "A Unified Course is the result of sharing experience and expertise in all facets of a course. It is a generalized course containing core topics as well as other optional topics. It also covers pedagogy, course contents, course supportive material and reference, assessments, and so on. The course adheres to national and international accreditation bodies while satisfying the intended learning outcomes that are directly aligned with program objectives." Working on the unified course is usually carried out during the planning and design phase of the project. A coordinator is assigned the task of putting together the final unified course, upload the results to the learning management system, and provide for future and continuous collaboration, maintenance, and updates. Figure 4 shows the unified course concept. Other instructors from different universities are also asked to upload their regular courses, and course material, thus at this point the system will consist of a unified course existing on a central location, along with a number of other institution-specific courses located in different spaces, connected all together through the Learning management system. New institutions who wish to join the network at this stage, can take advantage of the unified course to build their own specific courses, and at the same time may provide the network with feedback and suggestions to update the unified course. iJET ‒ Volume 9, Issue 8: "Learning in Networks", 2014 Figure 3. Collaborative network based on Moodle LMS Figure 4. Collaborative network for the production of unified programs Involved instructors from different universities can now enjoy a collaborative environment and can share teaching and research experiences. Students on the other hand, can communicate with their counterparts from within their institution and with other students from different universities. They can also communicate with other instructors, and get involved with lessons administered by different instructors. D. Collaboration from Student's Perspective Students are considered the main actors of the proposed learning system. Accordingly, special attention must be placed to provide students with an easy to use collaborative environment which can be useful and fun to use. Students can collaborate with fellow students from different universities using different activities. They can also collaborate with their instructors as well as other instructors from different universities. Students also have access to the unified courses and their resources. Figure 5 depicts student's access and collaboration. E. Collaboration from Teacher's Perspective Teachers have the opportunity to share their experience and teaching philosophy, not just on the course level, but 35 PAPER A FRAMEWORK FOR COLLABORATIVE NETWORKED LEARNING IN HIGHER EDUCATION: DESIGN & ANALYSIS also at the program level. They have access to the unified course as well as to other fellow teacher courses from different universities. Experienced teachers as well as newcomers can share teaching pedagogy, course text books, course outline, resources and references, and research ideas. Research groups amongst interested teachers can be easily formed, thus providing a fertile ground for future collaboration in research. Teachers can also stay in touch with their own students, as well as other students from different universities. F. Other issues related to design One of the strongest features of using Moodle LMS in the implementation of networked learning is its flexibility in controlling user access. The system permits administrators, teachers, and content creators to totally control who, what, when, and how others access their contents. Using this advanced feature of Moodle truly helps in organizing the most complex network. For any user to have access his/her username (profile) must exist within the system's database which can be accomplished in a variety of methods such as direct uploading, LDAP authentication, or online registration. Once a user profile exists within the system's database, it can be assigned any role by the administrator. In order to manage users and their roles within a complex network consisting of users from different universities, we view the user database as a hierarchy. V. ANALYSIS STAGE A. Social Network Analysis(SNA) The second part of the framework is the analysis of the participant’s interactivity that involve in the collaborative network learning within and among the institutions. This part intends to utilize the social network analysis technique that assists in understanding the interactivity of participants’ networked learning. SNA aims to identify the relationships based on how the actors are connected with each other [30, 31]. For example, Moore (1989) [32] identifies that student’s interactivity are three types of interactions in learning environments: learner-content, learnerinstructor, and learner-learner. We propose that SNA can help in representing and mapping participants’ relationships in the NL. This leads on to generating additional analytical data about the interactivity between the members of NL, and understanding behavior in NL environments. In our case, SNA is used to understand the flow of information, and the exchange of resources or activities among members of NL. Our primary focus in the analysis part is to study the exchange messages in the course activities and resources. SNA can be used to visualize and represent the social environment as a network based on the relationships by creating a graphical representation called a sociogram. The nodes of the sociogram represent the participants and the edges represent the connections between the nodes. The major analytical data from SNA calculates the cohesion of a network. The most important measures of SNA are “density” and “centrality. Density as a measure is a degree of the overall ‘connections’ between the participants. The density of a network is defined as the number of edges in a network divided by the maximum number of possible edges [30]. The value of density varies between 0 36 Figure 5. Collaboration Diagram showing a sample of two institutions, one program of study, and an instance of one course for each and 100% which means that the higher the density value is, the more interactions exist within the network [30, 33]. Centrality as a measure provides the participant that plays the central role in the network [31]. In our, case, this can reflects on the institution, course activity, resource, instructor, and student. This centrality can be done for each participant in the network by measuring the number “in-degree” and “outdegree” values. In-degree centrality counts the number of messages received to certain participant, whereas outdegree counts number of messages a participant has sent to other participants [33]. The SNA technique is used commonly in sociology and organizational studies, but it has been utilized in varies NL/CSCL research to study group interaction and, communication [34]. For example, Martinez, Dimitriadis, Rubia, Gomez, and de la Fuente (2003) [35] found that teacher’s presence has an effect on the network density. Daradoumis, Martinez-Mones, and Xhafa (2004) [36] used SNA to assess and identify the performance of the most effective virtual learning groups. B. Evaluation for collaborative learning network The evaluation stage of the collaborative learning network frameworks intends to quantitatively examine the relationships between perceived collaborative learning with participant’s satisfaction in different unified courses. For example, LaPoint and Gunawardena (2004) [37] denote that student satisfaction can be considered as an effective measure of learning outcomes. This will be implemented by preparing a questionnaire that will discover what participants' opinions and attitudes are towards the unified course contents, activities and resources. Participants will be people with various roles, including students, from different institutions involved in the networked learning. Results will be analyzed using different statistical techniques. VI. CONCLUSION The work presented in this paper has been guided by the authors' vision to enhance higher education in Jordan through collaborative work involving a number of local universities. The enhancement of higher education was translated into several goals and objectives to be satisfied http://www.i-jet.org PAPER A FRAMEWORK FOR COLLABORATIVE NETWORKED LEARNING IN HIGHER EDUCATION: DESIGN & ANALYSIS through the implementation of a system which is conceived, planned, and implemented as a result of shared knowledge, experience, expertise, and resources of participating universities. In order to arrive at a comprehensive system which successfully connects hundreds of stakeholders, starting from top policy makers down to the student level, the authors realize the need to convert the actual implementation steps into a general framework. The framework which is based in principle on previously studied learning theories provides interested researchers with general guidelines and procedures which can be easily customized to fit future learning projects. REFERENCES [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] World Bank, Education Reform for the Knowledge Economy II Project Information Document. World Bank, Washington, DC., 2009. G. Issa, SM. Hussain, H Al-Bahadili,"Competition-Based Learning: A Model for the Integration of Competitions with ProjectBased Learning using Open Source LMS, " International Journal of Information and Communication Technology Education (IJICTE),Vol. 10, No.1, IGI Publishing, 2014, pp.1 – 13. L.S. Vygotsky, Mind in Society. The Development of Higher Psychological Processes. M. Cole, V. John-Steiner, S. Scribner & E. Souberman, Eds. Cambridge, Massachusetts, Harvard University Press, 1978. D. R. Garrison, N. D. Vaughan. Blended learning in higher education: Framework, principles, and guidelines. John Wiley & Sons, 2008. D. R. Garrison, M. Cleveland-Innes, and T. S. Fung. "Exploring causal relationships among teaching, cognitive and social presence: Student perceptions of the community of inquiry framework." The Internet and Higher Education, vol 13, no.1, 2010, pp. 31-36. http://dx.doi.org/10.1016/j.iheduc.2009.10.002 A. Hope, "Factors for success in dual mode institutions." Commonwealth of Learning, Vancouver, Canada, 2006. R. T. Osguthorpe, and R. G. Charles. "Blended Learning Environments: Definitions and Directions." Quarterly Review of Distance Education vol. 4, no. 3, 2003, pp. 227-33. R. Huang, M. Ding, and H. Zhang. "Towards a design theory of blended learning curriculum." Hybrid Learning and Education, Springer Berlin Heidelberg, 2008, pp. 66-78. A. N. Leontiev, Actividade, consciência e personalidade. Buenos Aires: Ciências Del Honbre, 1978. P. Dillenbourg, "What do you mean by collaborative learning?." In Collaborative-learning: Cognitive and computational approaches, P. Dillenbourg, Eds., Oxford: Elsevier, 1999, pp. 1-19. R. Mitnik, M. Recabarren, M. Nussbaum, and A. Soto. "Collaborative robotic instruction: A graph teaching experience." Computers & Education, vol. 53, no. 2, 2009, pp. 330-342 http://dx.doi.org/10.1016/j.compedu.2009.02.010 M. M. Chiu,. "Flowing toward correct contributions during group problem solving: A statistical discourse analysis." the Learning Sciences vol. 17, no. 3, 2008, pp. 415-463. J. Roschelle, and S. Teasley, The construction of shared knowledge in collaborative problem solving, O'MALLEY C., Computer-Supported Collaborative Learning, 1995. C. Tu, and C. Michael, "Building active online interaction via a collaborative learning community." Computers in the Schools, vol. 20, no. 3, 2003, pp. 51-59. http://dx.doi.org/10.1300/ J025v20n03_07 C. Tu, Online collaborative learning communities: Twenty-one designs to building an online collaborative learning community. Libraries Unlimited, 2004. P. Vesely, B. Lisa, and S. John, "Key elements of building online community: Comparing faculty and student perceptions." MERLOT Journal of Online Learning and Teaching, vol. 3, no. 3, 2007, pp. 234-246. P. Shea, L. Chun, and P. Alexandra, "A study of teaching presence and student sense of learning community in fully online and webenhanced college courses." The Internet and Higher Education, vol. 9, no. 3, 2007, pp. 175-190. http://dx.doi.org/10.1016/ j.iheduc.2006.06.005 iJET ‒ Volume 9, Issue 8: "Learning in Networks", 2014 [18] P. Resta, and L. Thérèse, "Technology in support of collaborative learning." Educational Psychology Review, vol. 19, no. 1, 2007, pp. 65-83. http://dx.doi.org/10.1007/s10648-007-9042-7 [19] G. Stahl, K. Timothy, and S. Dan, Computer-supported collaborative learning: An historical perspective, Cambridge handbook of the learning sciences, 2006. [20] S. Banks, P. Goodyear, V. Hodgson, and D. McConnell, "Introduction to the special issue on Advances in Research on Networked Learning." Instructional Science, vol. 3, 2003, pp. 1-6. [21] C. Bereiter, "Design research for sustained innovation." Cognitive studies, bulletin of the Japanese cognitive science society, vol. 9, no. 3, 2002, pp. 321-327. [22] R. Sawyer, Social emergence: Societies as complex systems., Cambridge University Press, 2005. http://dx.doi.org/10.1017/ CBO9780511734892 [23] C. Jones, "Networked Learning: Stepping Beyond the Net Generation and Digital Natives.", In Exploring the Theory, Pedagogy and Practice of Networked Learning, Springer, 2012, pp. 27-41. http://dx.doi.org/10.1007/978-1-4614-0496-5_2 [24] A. Heylighen, et al., "Distributed (design) knowledge exchange.", Ai & Society, vol. 22, no. 2, 2007, pp. 145-154. http://dx.doi.org/10.1007/s00146-007-0122-5 [25] G. Siemens, Learning and knowing in networks: Changing roles for educators and designers., ITFORUM for Discussion, 2008. [26] S. Downes, "Places to go: Connectivism & connective knowledge." Innovate: Journal of Online Education, vol. 5, no.1, 2008. [27] A. Abran, J.W. Moore, Guide to the Software Engineering Body of Knowledge (SWEBOK®), IEEE Computer Society 2004 Guide, Angela Burgess, 2004. [28] B. Boehm, R. Turner, “Management Challenges to Implement Agile Processes in Traditional Development Organizations”, IEEE Software vol. 22, no. 5, 2005, pp. 30-39. http://dx.doi.org/10.1109/MS.2005.129 [29] Lee, Seiyoung; Yong, Hwan-Seung. Distributed agile: project management in a global environment. Empirical Software Engineer. 2010. [30] J.Scott, Social network analysis: A handbook. London: Sage, 1999. [31] S.Wasserman, and K. Faust, Social network analysis: Methods and applications. Cambridge: Cambridge University Press, 1997. [32] M. Moore, “Three Types of Interaction”, The American Journal of Distance Education, vol. 3, no. 2, pp. 1-7, 1989. http://dx.doi.org/10.1080/08923648909526659 [33] S.Borgatti, M. Everett, and L.Freeman, Ucinet for Windows: Software for Social Network Analysis. Harvard, MA: Analytic Technologies, 2002. [34] C. Haythornthwaite , “Exploring Multiplexity: Social Network Structures in a Computer-Supported Distance Learning Class”, The Information Society, vol. 17, 2001, pp. 211–226. http://dx.doi.org/10.1080/01972240152493065 [35] A. Martinez, Y., Dimitriadis, B., Rubia, E. Gomez, and P., de la Fuente, “Combining Qualitative Evaluation and Social Network Analysis for the Study of Classroom Social Interactions”, Computers & Education, vol 41, no. 4, 2003, pp. 353–368. http://dx.doi.org/10.1016/j.compedu.2003.06.001 [36] T. Daradoumis, T., Martinez-Mones, and F. Xhafa, “An Integrated Approach for Analyzing and Assessing the Performance of Virtual Learning Groups”, in Lecture notes in computer science, G. de Vreede, L. A. Guerrero, & G. M. Raventós, Eds, Berlin Heidelberg New York: Springer, 2004, pp. 289–304. [37] D. LaPointe, and C. Gunawardena, “Developing, Testing and Refining of a Model to Understand the Relationship Between Peer Interaction and Learning Outcomes in Computer-Mediated Conferencing”, Distance Education, vol, 25, no. 1, 2004, pp. 83-106. http://dx.doi.org/10.1080/0158791042000212477 AUTHORS Ghassan F. Issa, Haya A. El-Ghalayini, Ahmad F. Shubita, and Mohammed H. Abu-Arqoub are with University of Petra, Amman, Jordan. Submitted 22 April 2014. Published as re-submitted by the authors 26 May 2014. 37 Copyright of International Journal of Emerging Technologies in Learning is the property of International Journal of Emerging Technologies in Learning and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. J. EDUCATIONAL COMPUTING RESEARCH, Vol. 39(1) 17-36, 2008 USING SOCIAL NETWORK ANALYSIS TO UNDERSTAND SENSE OF COMMUNITY IN AN ONLINE LEARNING ENVIRONMENT DEMEI SHEN PIYANAN NUANKHIEO XINXIN HUANG University of Missouri-Columbia CHRISTOPHER AMELUNG Yale University JAMES LAFFEY University of Missouri-Columbia ABSTRACT This study uses social network analysis (SNA) in an innovative way to describe interaction and explain how interaction influences sense of community of students in online learning environments. The findings reveal differences on sense of community between two similarly structured online courses, and show unique interaction patterns for students in the two courses. These results should prove to be interesting to researchers of online learning both on a theoretical and methodological basis. INTRODUCTION The most popular format of distance learning today is online learning, which is defined as any class that offers its entire curriculum in the online course delivery mode, thereby allowing students to participate regardless of geographic location, independent of time and place (Richardson & Swan, 2003). Compared to 17 Ó 2008, Baywood Publishing Co., Inc. doi: 10.2190/EC.39.1.b http://baywood.com 18 / SHEN ET AL. traditional education formats, online learning is generally considered more costeffective and more convenient for learners. However, since online learning environments lack face-to-face communication among students and between students and instructors, online learners may experience feelings of isolation which can lead to dropping out (Cereijo, Young, & Wilhelm, 2001; Curry, 2000; Rovai & Wighting, 2005). Feelings of isolation have been related to a low sense of community (Haythornthwaite, Kazmer, Robins, & Shoemaker, 2000). Based on a qualitative study Eastmond (1995) argues that feelings of isolation can be alleviated when students work together in a community of learners and support one another. Understanding how a community of learners form and how social interaction may foster a sense of community in distance learning is important for both building theory about the social nature of online learning and for making practical recommendations to online instructors for methods to counter isolation. Rovai (2002b) demonstrated that building a sense of community can affect student satisfaction and learning. Interaction has been shown to be a critical element that influences sense of community in a learning context, because community members promote learning through dialog among members (Cunningham, 1992). Many researchers have examined interaction in online education and highlighted the importance of interaction for successful online learning (Palloff & Pratt, 1999; Shale & Garrison, 1990; Su, Bonk, Magjuka, Liu, & Lee, 2005). Most of the available research has employed interview and survey methods to study interaction, probing the perception of students; or employed content analysis to analyze posts in discussion forums. While these methods study individual students’ perspectives and expressions; there exists a different approach that studies the group or community through the interaction patterns among individuals so as to understand the types and nature of interactions that occur in online contexts. Social network analysis (SNA) provides a quantitative representation of the patterns of relationships among online members; it also provides visualization of patterns of interaction among members in networks, and enables quantitative comparisons between different networks or groups of people. The SNA approach offers a quantification and visualization of relations and interaction patterns that are not provided through survey or content analysis. This article describes the nature of interaction patterns in two online courses, which were similarly structured and were perceived by students to have different levels of sense of community. Given that prior survey and content analysis research establishes an association between student interaction and sense of community, the purpose of the current study is to determine if the ways that interaction is articulated in SNA are also associated with sense of community. Further the study examines the SNA representations to develop insight about the social nature of online learning and the potential value of using SNA to build theory and suggest improvements in online practices. The data are from surveys of students’ sense of community and log files detailing the interaction among SOCIAL NETWORK ANALYSIS / 19 the online students, instructor, and teaching assistant. The interaction type (read, post, and revision, etc.), the time of the interaction, the target object of the interaction, the actor, and the author of the target object were all recorded; in other words, the system captured the who, when, how, and where details of interaction. THEORETICAL FRAMEWORK Sense of Community and Interaction in Online Learning Sense of community is defined as “a feeling that members have of belonging, a feeling that members matter to one another and to the group, and a shared faith that members’ needs will be met through their commitment to be together” (McMillan & Chavis, 1986, p. 9). In online learning, Rovai (2002b) defines learning community as a social community of learners who share knowledge, values, and goals; he also concluded that sense of community in online learning is comprised of two components: 1) connectedness, which refers to students’ feelings of “cohesion, spirit, trust, and interdependence” (p. 325), and 2) learning, which is students’ feelings of “the extent to which their learning goals and expectations are satisfied.” The contribution of sense of community to online learning has been investigated by numerous researchers. Researchers have shown that sense of community in online learning environments contributes to student motivation and increased learning (Ertmer & Stepich, 2004; Palloff & Pratt, 1999; Wilson, 2001). Researchers have also shown that sense of community helps reduce feelings of isolation (Tinto, 1993) and facilitates learning by making more resources available in the form of other learners from whom learners can seek help (Frymier, 1993; Gibbs, 1995). Ertmer et al. (2004) also stated that establishing a sense of community is valuable because it helps provide the social context in which learning occurs. Sense of community is closely associated with interaction. According to Carabajal, Lapointe, and Gunawardena (2003), there are three dimensions of an online community: technological, task, and social; the social dimension refers to a sense of belonging, social-emotional bonds, and good relationships through frequent interaction during the online learning process. Tu and Corry (2002) propose a similar theoretical framework by stating that an eLearning community is comprised of instruction, social interaction, and technology. These frameworks of online learning indicate that interaction is important in a community. As summarized by Ertmer et al. (2004), Lipman (1991) identified five specific ways that interaction in a community is important and promotes learning: 1) listening to one another with respect; 2) trying to identify one another’s assumptions; 3) challenging one another to supply reasons for unsupported opinions; 4) building on one another’s ideas; and 5) assisting each other in drawing inferences from what was said. A study by Arbaugh (2000) confirms the relationship 20 / SHEN ET AL. between interaction and the sense of learning in a community. The study examined five Internet-based MBA courses and found that students’ perceptions of learning were associated with the instructor’s emphasis on interaction, ease of interaction, and classroom dynamics. Rourke, Anderson, Garrison, and Archer (2001) argue that perceived interaction with others is one of the cornerstones for the development of online learning communities. However, little empirical research exists that examines the ways interaction influences the formation and sustenance of sense of community in online learning environments. Furthermore, previous research has employed traditional method, for example, using students’ self-reported data and artifacts of communication. New methods such as social network analysis may be helpful in articulating interaction in online learning and the relationship between interaction and sense of community. Social Network Analysis SNA is a regularly used method within sociological and organizational studies for exploring human and social dynamics. However, there is a growing interest in using SNA in the field of education and online learning to determine information sharing patterns, understand social capital formation in networks, and implement program evaluation (e.g., de Laat, Lally, & Lipponen, 2004; Harrer, Zeini, & Pinkwart, 2006; Nurmela, Palonen, Lehtinen, & Hakkarainen, 2003). SNA is a method for representing and quantifying interaction which can augment the two approaches traditionally used by researchers to study interaction in online learning: content analysis (Anderson, 2003) and learner feedback (perception of students) (Rovai, 2002b). SNA is both a theoretical perspective and a methodology involving quantifying and visualizing observations of relational ties of members. There exist two distinct approaches to SNA, egocentric (ego network analysis) and sociocentric (complete network analysis). The egocentric approach focuses on individuals and is concerned with identifying and generalizing the relationships of core members; the sociocentric network analysis involves the quantification of relationships between people within a defined group, focusing on measuring the structural patterns of interaction between members, and consequently, how the patterns influence outcomes. In our study, we analyze data from the sociocentric approach. The measures of SNA in this article include: Network Density represents the actual number of ties in a network as a ratio of the total maximum ties that are possible with all the nodes of the network. A fully dense network has a network density value of 1 (above one when data contain edge weights, e.g., tie strengths), which indicates that all nodes are connected to each other. Since our analysis includes all the ties (weighted ties) between nodes the density values will often exceed 1. A network with a density value near 0 indicates that it is a sparsely-knit network. For an undirected graph with N nodes and M ties the density D is defined as D = 2M / N (–1). SOCIAL NETWORK ANALYSIS / 21 Centrality Degree is the number of ties to other actors in the network. It is measured by indegree and outdegree. Indegree indicates the amount of people that interact with a certain student. Outdegree is the amount of interaction a student initiates with others. Network Centralization is the degree of inequality or variance in the network as a percentage of that of a perfect star network of the same size. Measure of centralization is “an expression of how tightly the graph is organized around its most central point” (Scott, 2000, p. 89). METHODOLOGY Context of the Study and Participants The courses examined in this study were about the design of educational technologies and were taught to graduate students in a program for design and development professionals. Both of the courses taught design skills using similar methods and were fully online and delivered using the course management system Sakai. Both courses also had the same instructor and each had a teaching assistant that facilitated the course. There were 10 students in Course A who chose to participate and 15 participants in Course B. Table 1 reports student Table 1. Demographic Information of Participants Course A N (%) Course B N (%) Gender Male Female 3 (30%) 7 (70%) 7 (46.7%) 8 (53.3%) Age 21-25 26-30 31-35 Over 36 1 (10%) — 2 (20%) 7 (70%) 2 (13.3%) 3 (20.0%) 7 (46.7%) 3 (6.7%) Academic status Master Ph.D. Specialist Other 9 (90%) 1 (10%) — — 8 (53.3%) 5 (33.3%) 1 (6.7%) 1 (6.7%) 10 15 Demographic variables Total 22 / SHEN ET AL. demographic information. The participants in the two courses were adult learners, with an average age of more than 26. Only one participant in Course A and two in Course B reported their age as 21-25. Part I: Survey Study Procedure The students were given an online survey following each of three different class activities (units of the course which lasted approximately two weeks). The activities represent three different task structures: individual activity, group activity, and peer review activity. In the individual activity, students were required to complete a learning task individually, however, they could post to discussion boards asking questions or offering help to others. In the group activity, students were randomly assigned to teams of three members. They needed to communicate, negotiate and accomplish a task for which they would be graded as a team. In the peer activity, students worked as dyads to finish learning tasks. The dyads were randomly assigned by the instructor. The assignment called for dyad partners to review and offer feedback in a systematic fashion to their partners, but each student was graded separately for their own work. Each of the three activity types included social interaction via questions asked by the instructor of the students for response on the class discussion boards. The individual activity lasted two weeks, and the group and peer review activity each took three weeks. Students were asked to complete the survey and submit it electronically. All students were given the option of participating in this research project, and students who voluntarily completed the consent form became the sample of this study. Instruments The survey instrument measures sense of community with 6 items of the original 10 items assessing connectedness from the Classroom Community Scale (Rovai, 2002a). The Cronbach’s coefficient alpha was .93, indicating a high level of reliability. Students were asked to rate the level of agreement on a 7-point Likert scale where “1” represented strongly disagree and “7” represented strongly agree. The instrument was administered three times, immediately after each activity. RESULTS A two-way ANOVA was conducted using the individual courses and activity types (individual activity, group activity, and peer review activity) as independent variables and the final score of sense of community from the survey as the dependent variable. The results of the ANOVA are shown in Table 2. SOCIAL NETWORK ANALYSIS / 23 Table 2. Analysis of Variance for Sense of Community by Courses and Activity Type Source Sum of squares df Mean square F Partial eta squared p course_id 140.053 1 140.053 5.134* .076 < .05 47.382 2 23.691 .868 — > .05 .009 — > .05 activity_type activity_type * course_id Error .500 2 .250 1691.490 62 27.282 — Note: R squared = .099 (Adjusted R squared = .027) The overall R square is .099 (Adjusted R squared = .027). The main effect of course was statistically significant, F(1, 62) = 5.134, p < .05. Students in Course B had a significantly higher level of perceived sense of community than students in Course A. The main effect of activity type was not statistically significant. Part II: Social Network Analysis Part I analysis shows that students have a significantly different sense of community across the two classes and that this sense of community (in this context) does not vary between activities in the course even when they have different pedagogical structures. Describing the interaction patterns that characterize the two classes may reveal insights as to how sense of community can be understood and facilitated. Logs of all student activity in the system were automatically captured by the Context-aware Activity Notification System (Amelung, 2005). Interaction patterns and frequencies of interaction in the two courses were examined to articulate their relationship with sense of community. Since the two courses had a similar structure (each composed of similar objectives and activities), and were supervised by the same instructor, visualizing the interactions in activities in the two courses may help explain why students had different levels of sense of community in the two courses. SNA Graphs NetDraw 2.0 was used to create interaction diagrams and visualize interaction patterns for three activities in each course. The following three types of interactions were captured: • Content.read, which is reading a document posted in a shared folder in the system; 24 / SHEN ET AL. • Forum.read, which is reading posts from the instructor or other students in discussion forums in the system; • Forum.reply, which is replying to a message posted by another individual. The interactions between each individual were counted (see Table 3). The light lines indicate the unidirectional interactions, and the heavy lines indicate reciprocal interactions. Nodes with the same shape indicate partners for peer review activity. The diamond shape represents course instructor and the teaching assistants (TA). The Social network graphs attempt to represent the strength of social ties between nodes. SNA Measures The courses’ density, centralization, and centrality degree measure for each student were computed using UCINET 6.0, and the results are displayed in Tables 3 through 8. Network density indicates to what extent all nodes are connected to each other. Table 4 uses weighted ties to show that the density of all three activities were high when the instructors (instructor and TA) were included in the analysis; it decreased when excluding instructors which indicates that the instructor has a substantial impact on activity in the networks. Additionally Table 4 shows the densities of the two social activities in course B were higher than in course A when the impact of the instructor was removed from the network. Table 5 shows the means of indegree and outdegree when including the instructor and provides information about the activity level of members in their network. Indegree indicates that people have read a message from a certain person. Outdegree represents that a person has sent messages. To identify if there were significant differences on indegree and outdegree between Course A and Course B, two separate two-way ANOVA (Table 6) with course and activity as independent variables and outdegree and indegree as dependent variables were performed. The results indicate that participants in Course B had significantly higher levels of outdegree than the participants in Course A, F(1, 79) = 33.27, p < .01. Also, activity type had a significant effect on outdegree levels, F(2, 79) = 3.84, p < .05. Participants in peer activity had significantly higher levels of outbound interaction than in individual activity. In addition, there was a significant interaction effect of course and activity type. The outdegree in peer activity was the highest in Course B, while in Course A, it was the lowest. Similar results were found for indegree. There was a main effect of course, F(1, 79) = 6.03, p < .05, with participants in Course B having significantly higher levels of indegree than Course A. There was also a main effect of activity type, F(2, 79) = 3.30, p < .05. The indegree for the peer activity was significantly higher than individual activity. There was no interaction effect for course and activity type for indegree. Both Course A and Course B participants had similar patterns of indegree across activity types, for example, they both had the highest indegree SOCIAL NETWORK ANALYSIS / 25 in peer activity while the lowest indegree in the individual activity. And, when conducting ANOVA analysis for outdegree and indegree excluding the instructor, the results were similar for both of the two dependent variables. In addition, to further examine if the instructor influenced centrality degree, four sets of paired t-tests were performed to see if participants’ outdegree and indegree changed significantly with and without instructors in their network in Course A and Course B. Table 7 shows that the outdegree decreased significantly when instructors were excluded in both courses, t(35) = 5.99, p < .001, t(48) = 10.00, p < .001, respectively. For the measures of indegree, the difference was not significant between including instructor and excluding instructor in either course. Network centralization indicates to what extent a network is centered around or dominated by a few nodes (Scott, 2000). Table 8 shows that outdegree centralization in Course B was lower than Course A, meaning more people in Course B were active in communicating with others. Table 8 also shows the change of indegree centralization between with instructor and without instructor in course B was more dramatic than in Course A, which may indicate that students in Course B read the posts of the instructor more often than students in Course A. FINDINGS The ANOVA results on sense of community indicated that students doing similar activities in Course A and Course B had different levels of sense of community, and that differences on sense of community within a course across activity were not significant. Overall, both the visual graphs and measures of SNA demonstrated that students in Course B had more frequent interaction and more information exchange than students in Course A, which aligns with the statistical result that students in Course B perceived higher levels of sense of community than students in Course A. Visually, the SNA graphs of Course B for all activity types show more ties/lines/connections among members, with or without the instructor in the graph. The SNA graph shows that in Course B, all students are fully interconnected. However, in Course A there is one student that is not connected to other nodes in the network and is isolated from other students. The SNA measures convey a consistent story of interaction differences between the two courses. First, Course B had higher levels of network density than Course A. Also, the ANOVA results for centrality degree indicated that Course B had significantly higher levels of outdegree and indegree than Course A, showing that more interaction took place in Course B than in Course A (e.g., participants in Course B initiated and viewed more messages than those in Course A). In addition, the network centralization (Table 8) of the two courses indicates that the network of Course A is more focused around several essential nodes than Course B, implying that the overall interaction stretch of Course A is not as expanded as in Course B. 26 / SHEN ET AL. Table 3. SNA graphs for three activities (graphs for one activity were included). SOCIAL NETWORK ANALYSIS / 27 28 / SHEN ET AL. Table 4. The Density of the Interaction for Each Activity Course A Course B With instructor Without instructor With instructor Without instructor Individual activity 1.1479 0.4028 1.0842 0.1708 Group activity 1.7396 0.3889 1.5351 1.0500 Peer activity 0.9231 0.7179 2.6144 2.0875 Table 5. Descriptions of the Degree of Centrality Individual activity Group activity Peer activity Combined Outdegree Course A 15.50 (5.70) 23.42a (13.55) 11.50a (6.38) 16.81 (2.10) Course B 23.63b (8.99) 28.65 (14.98) 45.94b (18.03) 32.74b (1.80) Combined 20.14 (8.65) 26.48 (14.40) 31.18 (22.32) Indegree Course A 4.00c (2.09) 4.75 (5.05) 11.33c (5.13) 6.70d (4.96) Course B 2.56d (2.22) 30.35 (63.09) 35.31d (19.27) 22.74d (4.25) Combined 3.18c (2.24) 19.76 (49.49) 25.04c (19.07) Note: Means in the same row sharing the same letter superscript differ at p < .01. SOCIAL NETWORK ANALYSIS / 29 Table 6. ANOVA Result for Outdegree and Indegree by Courses and Activity Type Source Sum of squares df Mean square F Partial eta squared p 33.27 .30 < .01 Outdegree 5265.18 course 5265.18 1 activity_type 1216.29 2 608.14 3.84 .09 < .05 activity_type * course 3571.63 2 1785.82 11.29 .22 < .01 12501.49 79 158.25 Error Indegree course 5342.96 1 5342.96 6.03 .07 < .05 activity_type 5844.43 2 2922.21 3.30 .08 < .05 3169.95 2 1584.97 1.79 — > .05 69962.174 79 885.597 activity_type * course Error Note: R squared = .46 (Adjusted R squared = .42) for Outdegree analysis; R squared = .19 (Adjusted R squared = .13) for indegree analysis. The SNA graphs and measures show some similarities about interaction patterns across the two classes. First, the instructor played a pivotal role in both classes. The instructor provided information to get the work started and initiated interactions. The students tended to read what the instructor had posted and this is demonstrated by comparing the graphs of with and without instructor. Also, paired t-tests indicated that the outdegree for the two courses decreased significantly when removing instructor ties from calculations. This indicates that a fair amount of outbound interaction by participants was toward the instructors. Second, interaction patterns are influenced by the task types. For instance, students had substantially less reciprocal interactions in individual activity than in peer review activity in both classes, in a common sense way this makes sense, but it also suggests that research on interaction and community in online learning needs to pay more attention to the type of tasks given to students, as task type shapes activity. The role of task type and its impact on interaction in online learning needs further research. In addition, students’ interactions during peer review and group activities go beyond their assigned partners. They interacted with other classmates both accessing resources from others and providing resources to others. Third, the network centralization table shows that the networks of the two courses were not perfect star networks, which means interactions among students and instructor are not equally distributed; some students were more active than others; and this situation occurred in both classes. 30 / SHEN ET AL. Table 7. Paired T-Test Result on Degree Measures with Instructor and Without Instructor With instructor Without instructor Variables Mean Mean t df Sig. (2-tailed) Course A Outdegree (n = 36) 16.81 5.89 5.99 35 < .001 .69 5.89 1.70 35 > .05 Course B Outdegree (n = 49) 32.65 16.71 10.00 48 < .001 Course B Indegree (n = 49) 22.90 16.69 1.18 48 > .05 Course A Indegree (n = 36) T-test Further, the SNA shows differences in interaction patterns between the two classes. In Course A, students with high interaction scores generally are the ones that initiated interactions with other class members rather than waiting to be interacted with by others. While in Course B, students with high interaction scores generally are the ones that were interacted with by other students rather than those that initiated interactions. This may imply that students who had high interactions in both classes played different roles. Hypothetically in Course B, students who had high interactions might have been those that had influence on the flow of information in the class, that is, other students interacted with them because the quality of their work was superior to others or their postings/ideas were more meaningful than others. In other words, other students purposively selected to interact with them because of their role as a contributor to the social construction of the learning products. DISCUSSION The findings of the study show that interaction is strongly associated with students’ perceived sense of community in this online learning environment. The use of both traditional statistical analysis and innovative SNA techniques revealed that students in the course with higher interaction frequency perceived higher levels of sense of community, which supports the thesis that interaction plays a crucial role in forming students’ sense of community in the online course. This result is consistent with previous research, but the SNA representations extend 4.321% 5.615% 23.843% Individual activity Group activity Peer activity Outdegree 17.824% 32.076% 62.500% Indegree With instructor 33.333% 15.041% 20.248% 33.333% 15.041% 15.289% Indegree Without instructor Outdegree Course A 8.349% 6.185% 3.102% 38.286% 57.086% 45.346% 9.825% 6.133% 15.778% 12.111% 7.911% 8.667% Indegree Without instructor Outdegree Course B Indegree With instructor Outdegree Table 8. Network Centralization SOCIAL NETWORK ANALYSIS / 31 32 / SHEN ET AL. the previous findings to show how “ways” of interacting, not just “levels” of interaction have value for understanding interaction and how it relates to sense of community. The SNA vocabulary of ties, density, centrality and indegree and outdegree provide richness in describing interaction. This richness fills a void between purely qualitative analysis that has contextual grounding and simply quantitative measures that may not represent important distinctions within the findings. Our reflection on our results also leads us to speculate that perhaps SNA techniques can be better attuned to educational networks. For example, the sequence of activity or distinctions among roles could become part of the mathematical formulation or visual representation so as to provide additional and important information. In sum the findings suggest that by providing visual graphs and quantitative representations of the patterns and density of interaction, SNA helps articulate the way interaction unfolds in classes and shows potential to be a valuable method to study interaction in online learning environments. The findings have implications for understanding how interaction influences sense of community in online learning environments. First, with the SNA technique, the role of the instructor in fostering interaction in online learning environments can be parsed out. For instance the visualizations in graphs 2 and 4 in Table 3 and the statistics of Table 4 show how the density of the social experience is substantially formed around the instructor and that the instructor’s role has greater relative impact in Course A than in Course B. Second, while no significant differences were found for sense of community between different activities types, the SNA graphs and centrality degree measures indicate that the frequency of interaction for peer review and group activity is higher than individual activity. One explanation may be that the sample size was too small to highlight differences, but another explanation may lie in further examination of the types of connections between members, such as in the differences between high interaction students in courses A and B. Further studies with larger sample sizes and more course activities may provide explanations for why different patterns across courses are significant but not within courses. Third, SNA graphs show that some students were left out of the communication network, which influenced the information flow. These three implications suggest that more explicit monitoring of how students attend to instructor communication may lead to better online learning processes; that collaborative activities provide students more opportunities and obligations to communicate and work together and thus learn from one another; and that instructors need help for early identification of students who are lagging in their participation. SNA visualizations may be helpful for educators to the extent they can become part of the interface for online learning to provide a dynamic overview of class interaction such as are some students dominating and others lurking. These visualizations may provide instructors with quick feedback on instructional efforts to further engage students in the class, to get more even distribution of SOCIAL NETWORK ANALYSIS / 33 student initiative and breadth of who talks to whom, as well as providing feedback on efforts to promote different forms of information flow and interaction in the class. In addition to visual representations which may help the instructor see what is going on, the SNA metrics may be useful for providing indicators relative to standards for class interaction that an instructor wants to maintain. For example, an instructor might target a density of 2.0 for a particular unit. Of course these forms of metrics for pedagogy seem foreign to us now, but as we implement tools with more precision for understanding social interaction we may find new ways of assessing instructional practices. The results of this study should prove interesting to researchers of online learning both on a theoretical and methodological basis. The construct of sense of community has been shown by other researchers to be a replicable and important element in online education, and interaction is proving to be a key influence in promoting sense of community. The current study supports and extends previous research results by examining a new learning context and using new measures. Further, the results of our study demonstrate and provide support for social network analysis as a methodology for understanding interaction and its contribution to sense of community. As a powerful analytic and visual tool, SNA helps reveal the structural characteristics of the group, identifying the connections and possible disconnects among participants in online courses. In addition, rather than being limited to one form of representation for activity, the data from the present study represents multiple ways to assess the digital experience of learners. This type of examination is made possible by new tools such as SNA for doing the analysis, but also tools for monitoring behavior in online environments such as the Context-aware Activity Notification System (Amelung, 2005) that traced and recorded all interactions in the online environment. An interesting aspect of our findings is that we characterize the two courses as similar because they have the same instructor applying a common structure and instructional approach to similar course topics (learning to design) for similar students (graduate students in the same program). However the students’ interaction patterns were substantially different. What could cause this difference? One possible cause may be that early on in the course different events occurred that started the students behaving differently. For example, perhaps a discussion in one course was particularly engaging and viewed by students as empowering, whereas in the other class a discussion may have led to acrimony among students. Another factor that may have contributed to the different interaction patterns could be individual differences. While the students were all graduate students in the same program, students in one course may have had higher motivation or been more socially dependent than those in the other course. While SNA does not show us why students in one course behaved differently than the other, the results of indegree and outdegree analysis show that the instructor was more effective in Course B in getting nearly all students to interact with one another at a relatively high level. Although a number of students in Course A similarly interacted at a 34 / SHEN ET AL. high level, the density and type of interactions differed between the courses. Did the instructor fail to address some problems in Course A or was he particularly brilliant in Course B? Unfortunately SNA analysis cannot answer these questions. This complexity of the findings suggests that future studies may want to include content analysis of interactions and some measures of individual differences with SNA to provide more comprehensive understanding of the social nature of the online learning experience. Finally, although the quantitative analysis and SNA together provide an innovative way to examine interaction and sense of community in online learning environment, there are limitations. The data used in the study captured online interaction in the course management system; however, students may communicate using other tools outside of the course management system, such as e-mail. Therefore, the data may not be comprehensive in representing all the interaction of students. Future studies may benefit from efforts to capture information about communication that happens outside of the course tools and by combining SNA approaches with traditional approaches of surveys and content analysis. REFERENCES Amelung, C. (2005). A context-aware notification framework for developers of computer supported collaborative environments. Unpublished doctoral dissertation, University of Missouri-Columbia, Columbia. Anderson, T. (2003). Getting the mix right again: An updated and theoretical rationale for interaction. International Review of Research in Open and Distance Learning, 4(2), 1-14. Arbaugh, J. B. (2000). How classroom environment and student engagement affect learning in Internet-based MBA courses. Business Communication Quarterly, 63(4), 9-26. Carabajal, K., Lapointe, D., & Gunawardena, C. N. (2003). Group development in online learning communities. In M. G. Moore & W. G. Anderson (Eds.), Handbook of distance education (pp. 217-234). Mahwah, NJ: Lawrence Erlbaum Associates. Cereijo, M. V. P., Young, J., & Wilhelm, R. W. (2001). Factors facilitating student participation in asynchronous web-based courses. The Journal of Computing in Teacher Education, 18(1), 32-39. Cunningham, R. D. (1992). Beyond educational psychology: Steps toward an educational semiotic. Educational Psychology Review, 4, 165-194. Curry, D. B. (2000). Collaborative, connected and experiential learning: Reflections of an online learner. Retrieved Nov. 2004, from http://www.mtsu.edu/~itconf/proceed01.2.html de Laat, M. F., Lally, V., & Lipponen, L. (2004). Patterns of interaction in a networked learning community. In ELearn: World Conference on E-Learning in Corporate, Government, Healthcare, and Higher Education, Washington DC, USA, 1-5 Nov 2004. AACE. Eastmond, D. V. (1995). Alone but together: Adult distance study through computer conferencing. Cresskill, NJ: Hampton Press, Inc. SOCIAL NETWORK ANALYSIS / 35 Ertmer, P. A., & Stepich, D. A. (2004). Examining the relationship between higher-order learning and students’ perceived sense of community in an online learning environment. Paper presented at the proceedings of the 10th Australian World Wide Web conference, Gold Coast, Australia. Frymier, A. B. (1993). The impact of teacher immediacy on students’ motivation over the course of a semester. Paper presented at the annual meeting of the Speech Communication Association, Miami Beach, FL. Gibbs, J. (1995). Tribes. Sausalito, CA: Center Source Systems. Harrer, A., Zeini, S., & Pinkwart, N. (2006). Evaluation of communication in websupported learning communities—An analysis with triangulation research design. International Journal of Web Based Communities, 2(4), 428-446. Haythornthwaite, C., Kazmer, M., Robins, J., & Shoemaker, S. (2000). Making connections: Community among computer-supported distance learners. Paper presented at the Association for Library and Information Science Education 2000 Conference. San Antonio, Texas. Retrieved July 3, 2003 from: http://www.alise.org/conferences/conf00_Haythornthwaite_Making.htm Lipman, M. (1991). Thinking in education. New York: Cambridge University Press. McMillan, D. W., & Chavis, D. M. (1986). Sense of community: A definition and theory. Journal of Community Psychology, 14(1), 6-23. Nurmela, K., Palonen, T., Lehtinen, E., & Hakkarainen, K. (2003). Developing tools for analyzing cscl process. In B. Wasson, S. Ludvigsen, & U. Hoppe (Eds.), Designing for change in networked learning environments. Proceedings of the International Conference on Computer Support for Collaborative Learning 2003 (pp. 333-342). Kluwer, Dordrecht. Palloff, R. M., & Pratt, K. (1999). Building learning communities in cyberspace. San Francisco, CA: Jossey-Bass. Richardson, J. C., & Swan, K. (2003). Examine social presence in online course in relation to students’ perceived learning and satisfaction. Journal of Asynchronous Learning Networks, 7(1), 68-88. Rourke, L., Anderson, T., Garrison, D. R., & Archer, W. (2001). Assessing social presence in asynchronous text-based computer conferencing, Journal of Distance Education, 14(3), 51-71. Rovai, A. P. (2002a). Development of an instrument to measure classroom community. Internet and Higher Education, 5(3), 197-211. Rovai, A. P. (2002b). Sense of community, perceived cognitive learning, and persistence in asynchronous learning networks. Internet and Higher Education, 5(4), 319-332. Rovai, A. P., & Wighting M. J. (2005). Feelings of alienation and community among higher education students in a virtual classroom. The Internet and Higher Education, 8(2), 97-110. Scott, J. (2000). Social network analysis: A handbook. London: Sage. Shale, D., & Garrison, D. R. (1990). Introduction. In D. G. D. R. Shale (Ed.), Education at a distance (pp. 1-6). Malabar, FL: Robert E. Kriger. Su, B., Bonk, C. J., Magjuka, R. J., Liu, X., & Lee, S. (2005). The importance of interaction in web-based education: A program-level case study of online MBA courses. Journal of Interactive Online Learning, 4(1), 1-19. Tinto, V. (1993). Leaving college: Rethinking the causes and cures of student attrition (2nd ed.). Chicago: University of Chicago Press. 36 / SHEN ET AL. Tu, C., & Corry, M. (2002). E-Learning communities. The Quarterly Review of Distance Education, 3(2), 207-218. Wilson, B. G. (2001). Sense of community as a valued outcome for electronic courses, cohorts, and programs. Retrieved October 2005 at http://carbon.cudenver.edu/~bwilson/SenseOfCommunity.html Direct reprint requests to: Dr. Demei Shen School of Information Science and Learning Technologies University of Missouri-Columbia 303 Townsend Hall Columbia, MO 65201 e-mail: dmsry6@mizzou.edu http://mym.cdn.laureatemedia.com/2dett4d/Walden/EDUC/8345/CH/mm/anatomy_of_elearning/index.html Delivery- online learning(click on)
Purchase answer to see full attachment