This research pursues answers for the following main research questions. What ways can be applied for group formation that tailors to individual students’ characteristics and fits well into the existing collaborative learning environments? What problems exist widely in group collaboration and what are the factors that may lead to these problems? What approach can be adopted for automatically diagnosing these identified group collaboration problems in a collaborative learning environment?
1.1 Problem Statement and Motivation
Collaborative learning enables individual students to combine their own expertise, experience and ability to attain a mutual learning goal. Teachers are key performers in the process of structuring and managing online group work in current collaborative learning environments (CLEs). There is limited or no support for them to cope with tasks relating to organizing effective groups that satisfy individual students’ needs and assessing their problems in the collaborative process. As a result of this, the teachers must adopt a direct-manipulation method of interaction to cope with these tasks [1]. However, this direct-manipulation method is very time-consuming and labour-intensive for information gathering, retrieving and filtering. Along with the increasing use of online collaboration, there is a growing need to enhance the delivery methods and to simplify how teachers manage collaborative group work. The composition of groups is one of the factors that determine the effectiveness of collaborative group work, and is affected by several variables, including the demographics of the group members such as age, gender and race, the size of the group, and other differences between participants [2], and the allocation of students to such groups should take those factors into consideration. Furthermore, for a group to function effectively in a given learning environment, teachers should identify specific student characteristics and the group type (homogeneous or heterogeneous) which they understand to be appropriate for the learning activity [3]. However, the approaches adopted by teachers for group formation are usually forming ad-hoc groups in which these two aspects are ignored. Take self-created groups and computational randomly assigned groups as an example, it could be argued that these approaches provide no particular educational benefits. Self-created groups in particular friendship groups usually tend to avoid heterogeneity [4]. Randomly assigned groups do not ensure that students satisfy their individual needs. Recent work has highlighted how consideration of learning styles in the process of group formation for collaborative learning can have a positive impact[5,6,7]. However, current research does not suggest an approach that can automatically and efficiently form learning style groups. It motivates this research to propose an approach for group formation based on students’ learning styles. Some recent empirical studies including [8,9,10,11,12,13,14] have revealed that there is still a variety of problems existing in group collaboration, which eventually affects the effectiveness of collaborative group work. Some problems are caused by factors not directly related to the students such as challenges inherent in virtual communication relying solely on written language, insufficient and ambiguous instructions, and problems presented by working in different time zones. These studies also indicate that student-induced problems are the most serious. However, current literature does not systematically address the major student-induced group collaboration problems and the factors that may cause such problems. This issue motivates the research carried out in this thesis to identify student perceptions of the major group collaboration problems and their causes. Assessing these collaboration problems can assist teachers or moderators to understand and evaluate how individual students perform in a collaborative group as well as help students to reflect on their own learning. However, judging the existence of these problems is a complex task because a variety of such problems exist and distinct methods or techniques are required to support the analysis of these problems. Current applications that support online collaboration (including single tools such as forums and wikis as well as collections of tools such as collaborative learning environments) have limited or no support for monitoring the collaborative process and thus assessing the problems encountered by individual students and groups.
A number of research studies in interaction analysis for collaborative activities including [9,30,31,33,91,164,166] have indicated that quantitative data relating to student interactions with a collaborative learning system can account for the behaviours of individual students and collaborative groups. For example, Talavera and Gaudioso suggested that the number of threads started by an individual student can indicate the degree of involvement to produce a contribution [15]. Therefore, this research also seeks to propose an approach that can automatically diagnose the identified types of group collaboration problems based on student interactions with a collaborative learning environment. This aspect motivates this research to explore a multi-agent architecture for supporting online collaborative learning.
1.2 Research Aim and Objectives
The aim of this research is to explore solutions for improving the delivery of support for group work in collaborative learning environments, which can provide an enhanced and efficient way for teachers to cope with tasks of constructing collaborative groups and diagnosing group collaboration problems.
To achieve this aim, the following objectives are to be addressed.
- Identify major student-induced group collaboration problems and their causes from the perspectives of students via a nationwide survey, and provide a machine-readable form of the linkages between the problems and their causes.
- Propose an approach that can automatically diagnose the identified types of group collaboration problems based on student interactions with a collaborative learning environment.
- Carry out an evaluation of the diagnostic mechanism developed using a mixture of methods to determine its validity and effectiveness in ascertaining the existence of the collaboration problems identified on a test dataset.
The contributions of this thesis are presented as follows. - Based on the results from a survey-based study, this research identifies major student-induced group collaboration problems and their causes from the perspectives of students, and establishes a machine-readable representation of the linkages between the major problems and their causes. This is the first study that systematically addresses this issue, and the representation provides a unique perspective on the linkages between the problems and causes identified.
1.3 Definition of some key words
Artificial Intelligence: is the branch of computer science concerned with making computers behave like humans. The term was coined in 1956 by John McCarthy at the Massachusetts Institute of Technology. Artificial intelligence includes the following areas of specialization:
- Games Playing: Programming computers to play games against human opponents
- Expert system: Programming computers to make decisions in real-life situations (for example, some expert systems help doctors diagnose diseases based on symptoms)
- Natural language: Programming computers to understand natural human languages
- Neural network: Systems that stimulate intelligence by attempting to reproduce the types of physical connections that occur in animal brains
- Robotics: programming computers to see and hear and react to other sensory stimuli
The delivery of intelligent support for group work is a complex issue in collaborative learning environments. This particularly pertains to the construction of effective groups and assessment of collaboration problems.
Collaborative learning: is a situation in which two or more people learn or attempt to learn something together, unlike individual learning, people engaged in collaborative learning capitalize on one another’s resources and skills(asking one another for information, evaluating one another’s ideas, monitoring one another’s work, etc.) and also collaborative learning refers to methodologies and environments in which learners engage in a common task where each individuals depends on and is accountable to each other. These include both face-to-face conversations and computer discussions (online forums, chat rooms etc.). Methods for examining collaborative learning processes include conversation analysis and statistical discourse analysis.
Extensible Markup Language: XML code, a formal recommendation from the World Wide Web Consortium (W3C), is similar to Hypertext Markup Language (HTML). Both XML and HTML contain markup symbols to describe page or file contents. HTML code describes Web page content (mainly text and graphic images) only in terms of how it is to be displayed and interacted with. XML data is known as self-describing or self-defining, meaning that the structure of the data is embedded with the data, thus when the data arrives there is no need to pre-build the structure to store the data; it is dynamically understood within the XML. The XML format can be used by any individual or group of individuals or companies that want to share information in a consistent way. XML is actually a simpler and easier-to-use subset of the Standard Generalized Markup Language (SGML), which is the standard to create a document structure.
1.4 Concept of Artificial intelligence.
Artificial Intelligence (AI) is a branch of Science which deals with helping machines finding solutions to complex problems in a more human-like fashion. This generally involves borrowing characteristics from human intelligence, and applying them as algorithms in a computer friendly way.
AI currently encompasses a huge variety of subfields, from general-purpose areas such as perception and logical reasoning, to specific tasks such as playing chess, proving mathematical theorems, writing poetry, and diagnosing diseases. Often, scientists in other fields move gradually into artificial intelligence, where they find the tools and vocabulary to systematize and automate the intellectual tasks on which they have been working all their lives. Similarly, workers in AI can choose to apply their methods to any area of human intellectual endeavor. In this sense, it is truly a universal field.
AI is the computer-based exploration of methods for solving challenging tasks that have traditionally depended on people for solution. Such tasks include complex logical inference, diagnosis, and visual recognition, comprehension of natural language, game playing, explanation, and planning.
AI is the study of how to make computers do things which at the moment people do better. This is ephemeral as it refers to the current state of computer science and it excludes a major area; problems that cannot be solved well either by computers or by people at the moment.
Having addressed the aims and objectives of this thesis in Chapter 1, this chapter presents a review of relevant literature which provides a theoretical foundation of the thesis. The emphasis of this literature review has been laid on current delivery of support for group work in collaborative learning environments with regard to the formation and diagnosis for groups, and theories and practice relating to the topics of interest. This literature review has identified the gaps in research that motivate this thesis.
2.1 Collaborative Learning Environments
A collaborative learning environment (CLE) is a web-based educational system that provides collaborative learning specific functionalities (i.e. structuring and managing the collaboration [16]) as well as other supporting functionalities for online learning (e.g. designing, managing and delivering learning content).
This specific study focused on face-to-face, formal, peer collaborative learning. Most studies on collaborative learning take constructivism, as the theoretical underpinning of peer collaborative learning, because they focus on making meaning through social interactions. Scholars of collaborative learning believed that collaborative learning can help students accomplish tasks that cannot be accomplished individually, by leveraging knowledge, skills, and resources between participants, as well as creating circumstances for participants to help each other and con-construct knowledge. Although most studies have proved the positive effects of collaborative learning, not all collaborative learning activities were successful.
Collaboration at the conceptual level, involves:
⢠Awareness ┠We become part of a working entity with a shared purpose
⢠Motivation ┠We drive to gain consensus in problem solving or development
⢠Self-synchronization ┠We decide as individuals when things need to happen
⢠Participation ┠We participate in collaboration and we expect others to participate
⢠Mediation ┠We negotiate and we collaborate together and find a middle point
⢠Reciprocity ┠We share and we expect sharing in return through reciprocity
⢠Reflection ┠We think and we consider alternatives
⢠Engagement ┠We proactively engage rather than wait and see
Fig: 1
Dimensions along which collaboration can be structured include but are not limited to the allocation of members to groups [17], assigning group members to roles such as =producer’ and =reviewer’, and regulating who can interact with whom over time [18]. Forms through which collaboration can be managed include collecting interaction data, constructing models of interaction, comparing with desired state, moderating [16], etc. The supporting functionalities constitute the basic platform for online collaborative learning as for other e-learning forms, which include administration, content management, the learning workplace and tools for interaction (e.g. chats, forums, bulletin-boards) [19]. This definition of a CLE is also used by prior research in web-based collaborative learning environments including [20,21,22]. Current collaborative learning environments, which are better known as Learning Management Systems (LMSs), are used to distribute courses over the Internet with features for online collaboration. Some common LMSs are Moodle [23], LAMS [24] and Blackboard Learn [25]. The main feature of the existing collaborative learning environments is that they consist of courses that contain activities and resources. Students can take an online course by participating in the activities arranged for the course. Here an activity means the work to be completed by students for the purpose of learning or assessment. There are mainly three common types of activities that are supported by these collaborative learning environments: informative activities (e.g. noticeboards, announcements, and sharing resources), collaborative activities (e.g. chats, forums, and wikis), and assessment activities (e.g. choices, questions and answers, and submitting files). An activity-based collaborative learning environment is structured mainly for designing, delivering and managing such activities. Such an environment supports various functionalities: content management allows various activities to be defined and arranged for a particular course; tools for supporting activities provide different ways to present theactivities; a collaborative workplace allows online students to carry out learning activities together and interact with each other remotely in synchronous or asynchronous ways; administration allows technicians to maintain the collaborative learning system and course managers to manage online courses. Moodle is a typical activity-based collaborative learning environment, which is suitable both for individual learning and collaborative learning. Moodle offers 13 different types of activities that a student can complete via interacting with other students and/or the teacher. Among these activities, there are five types of activities that are used to support group work: chats, forums, wikis, blogs and glossaries (i.e. lists of definitions that can be created and maintained collaboratively). Moodle also supports a range of resource types such as files, pages (in HTML format) and links (URLs) that a teacher can add to a course to support student learning. It possesses all the supporting functionalities that are needed for a CLE. Regarding the collaborative learning specific functionalities, Moodle provides a basic level of support for structuring and managing collaboration. These include manually created or randomly created groups, logs of student participations (view, add, update and delete) in any group activities and reports on the number of the hits on the group activities. LAMS is an authoring and delivering system for online collaborative learning activities. LAMS is different from Moodle in that it is capable of capturing sequences of learning activities which involve groups of students, rather than a single activity or simply content. This is because LAMS has been developed in accordance with the Learning Design approach [36], which emphasises the capturing of the process of education instead of simply content. It is different from the well developed approach for e-learning which is dedicated to the authoring of content-based, self-paced learning objects for individual learning. LAMS provides 22 various types of activities of which six are for the group activities: chats, web conferencing, forums, Google maps (to add students’ own place markers and view others’ markers), mind maps and wikis. LAMS also allows a teacher to add a number of resource types such as files, URL links and zipped websites into a sequence of learning activities. Similar to Moodle, LAMS incorporates all the required supporting functionalities for a CLE and some simple collaborative learning specific functionalities such as teacher or student selected groups and random-created groups, and logs of user access to the group activities (for example, viewing a forum or a thread). Blackboard Learn is a commercial application while Moodle and LAMS are open source LMSs. It provides a broad range of activities and resource types that a teacher or student can add to a course for supporting teaching and learning. There are six types of activities that deliver the support for group work including forums, glossaries, chats, web conferencing, blogs and wikis. Blackboard Learn comprises all the supporting functionalities that are needed for a CLE but provides very limited support for structuring and managing collaboration. Forms of support includes student self-enrolled or teacher manually enrolled groups and randomly created groups, and logs of user accesses to the group activities. Although providing simple methods for group formation, Blackboard Learn does not support the association of created groups with any group activities as Moodle and LAMS provide. Blackboard Learn can only create groups for certain activities which include discussion boards, email lists, file exchanges and online chats. As can be seen from the above discussions, the existing collaborative learning environments provide a variety of supporting functionalities for online collaborative learning. Although they provide support for teachers to create collaborative groups, the methods adopted for constructing groups do not tailor to individual students characteristics because students are usually assigned to groups manually by teachers or randomly by the systems. Moreover, most contemporary CLEs are capable of tracking the accesses and actions performed by users, which enable student interactions stemming from the collaborative activities examined to be captured. However, they do not provide support for teachers to check the progress of student collaboration and thus to assess the collaboration problems. To investigate these issues, the following sections present a discussion of the methods, tools, theories and practice for group formation, prior research which revealed the problems impeding group collaboration, and the methodologies for establishing the diagnostic methods.
2.2 Collaborative Learning Problems
Collaborative learning enables individual students to combine their own expertise, experience and ability to accomplish a mutual learning goal. Teachers are key performers in the process of structuring and managing online group work in current collaborative learning environments (CLEs). There is limited or no support for them to cope with tasks relating to organizing effective groups that satisfy individual students’ needs and assessing their problems in the collaborative process. As a result of this, the teachers must adopt a direct-manipulation method of interaction to cope with these tasks.
The composition of groups is one of the factors that determine the effectiveness of collaborative group work, and is affected by several variables, including the demographics of the group members such as age, gender and race, the size of the group, and other differences between participants, and the allocation of students to such groups should take those factors into consideration.
Through a thorough review of literature, three major categories of group collaboration problems have been identified including
- Poor motivation,
- Lack of individual accountability and
Negative interdependence.
This review also indicates that there are several problem scenarios existing which can reveal the same category of group collaboration problem. Next, a brief summary of the problem scenarios identified regarding each category of group collaboration problem is provided.
Concerning =poor motivation’, two problem scenarios were identified. The first describes a scenario in which all the members in a collaborative group could post with an asynchronous collaboration tool (e.g. forums, wikis and blogs) to discuss a given learning topic and a student in the group made a post irrelevant to the learning topic. The second denotes a situation in which the members in a collaborative group were expected to provide in-depth reflective responses to a on a given learning topic or material and one of the group members made a post that contained several grammatical and/or spelling errors which was difficult to understand. These two scenarios were revealed from several studies in online group work. Regarding =lack of individual accountability’, three problem scenarios were recognized. The first represents a situation in which the members of a collaborative group discussed online to accomplish a piece of group work with an asynchronous collaboration tool and an individual student hadn’t contributed much during the online discussions. The second scenario describes a situation in which a deadline was set for a piece of group work and the members needed to complete the work together (no role division within the group), and one member was negligent in meeting the deadline. Furthermore, the final scenario can be explained as that each member in a collaborative group was allocated with a role to complete the group work and one student did not complete his or her assigned work. These problem scenarios were identified from research regarding the common collaboration problems that were faced by students participating in an online group project. In terms of =negative interdependenceâ, two problem scenarios were identified. One depicts a situation in which a collaborative group was assigned a piece of group work and all the members were desired to discuss the solutions together; however, they had given little feedback to each other about each otherâs thoughts. The other denotes a situation where the workload of a collaborative group was not shared fairly; one student in the group had made most of the work and other members did little or no work. These two scenarios reveal the problems possessed by individual groups whose members have negative relationships with regard to collaboration. The first scenario was identified from which noted that limited student participation in online discussion appears to be a persistent problem. The second scenario is known as the ‘free-rider’ problem identified by Roberts and McInnerney as one of the common problems of online group learning. Other studies that also noted these two problem scenarios include.
As can be seen from the above discussion, a total of seven problem scenarios corresponding to several sub-categories of group collaboration problems were identified from the literature. The survey presented in this chapter addressed the seven problem scenarios and the factors that may cause these problems.
The structure of the remaining sections is organized as below. Section 5.2 presents the methodology that was applied for conducting the survey-based study.
This includes several aspects: (i) the general research design including the type and scope of survey that was used, the administration of the survey and the ethical consent for this project; (ii) the targets of the survey and the method of inviting the participants to take the survey; (iii) the data collection instrument and procedure adopted; (iv) and the data analysis techniques that were applied. Following that, Section 5.3 presents the results that were obtained from the survey. Four aspects of results are presented. First, Section 5.3.1 summarizes the demographic information about the respondents. Second, Section 5.3.2 describes students’ views on the seven problem scenarios (i.e. whether they have experienced the problems or not) and what factors can cause such problems. Finally, Section 5.3.4 analyses the popularity of various asynchronous collaboration tools that the students had previously used for completing the group work. Section 5.4 discusses how the set of major group collaboration problems and their causes were determined from the survey results. In Section 5.5, a detailed description of the XML-based representation is provided, which includes the motivation for adopting XML for the representation, the hierarchical structure of the XML elements, a code fragment of the XML representation and the validation of the XML created. This section also discusses the potential applications of the XML-based representation. In addition, a summary of this chapter is provided in Section 5.6.