Frank Kurzel, Lecturer, School of Communication, Information and New Media, University of South Australia, Magill, SA 5072. Email: Frank.Kurzel@unisa.edu.au
Dr. Jill Slay, Senior Lecturer, School of Computer and Information Science, University of South Australia, MAWSON LAKES, SA 5095, Australia. Email: Jill.Slay@unisa.edu.au
In this paper, we describe the techniques we plan to overlay on an adaptive multimedia learning environment (AMLE) to provide alternate instructional methodologies to the learner. The system supports the WWW as its addressing space but uses the local client areas to store media items expensive in terms of delivery time. It provides link annotations in its adaptivity and employs an overlay student model with stereotyping when accessing the course content. The structure of the domain is based on concepts that can be presented in a variety of ways; multimedia players display the content on the client computer.
Computers have been used in the educational process since the early sixties with early applications generally improving in their ease of use and accommodation of more media rich items. The World Wide Web (WWW) has been introduced progressively into educational environments to support teaching. Typically, educational multimedia artefacts are now placed within some Learning Environment (LE) that is distributed over the WWW and accessible through some browser.
Online course materials have supported traditional lectures and tutorials; some have replaced them. However these and many of the newer systems that have been developed relied on low-levels of interaction and have failed to use features that has become available, often failing to maximise the potential of current multimedia resources.
Maurer (Maurer 2002) argues that even though the handling of different media items has improved over the years, the major improvements have come from other areas. For example, the WWW has provided us with a distribution infrastructure that overcomes distance and time, and hypertext/hypermedia environments provide associative access to educational materials. Embedded within these online educational materials have been multimedia artefacts that it is proposed, provide a possible enhancement to the presentation of on-line materials.
However, more fundamental problems exist with the current educational systems; generally, the coarse grained nature of online materials limits reusability and the subsequent degree of personalisation that can occur at the content level. In particular, systems don’t take account of prior knowledge, nor do they account for any differences in learning style. More significant is the inability to alter the learning theory that might be employed in the delivery of the course by either the instructor, or the learner.
Within this paper, we describe the design goals and decisions undertaken in the development of a prototype of a multimedia learning environment for the presentation of course materials; we call this system an Adaptive Multimedia Learning Environment (AMLE). The advantages that this system will provide will include the greater flexibility that students will enjoy in interacting with an instructional environment rich in different media items and tools for their perusal.
It has been argued (Eklund 1998) that Adaptive Hypermedia (AH) has the potential to individualise instruction in higher education. Interbook (Brusilovsky, Schwarz et al. 1996) demonstrated personalising features that could account for individual differences in knowledge. Brusilovsky used the term concept when referring to the elementary pieces of knowledge with a domain. This notion of a concept as a fundamental unit has subsequently been employed in instructional systems.
There is also the recognition that ‘multimedia has enhanced the cognitive flexibilty of computer based learning through its ability to restructure knowledge presentations to meet changing situations’ (Eklund 1998). Multimedia objects have been placed within learning environments with positive results (Bonkhorst 1998; Herrington and Oliver 1999; Looi and Ang 2000; Ricketts, Wolfe et al. 2000; Gibson 2002) to satisfy particular roles in the learning process. Some media formats may not satisfy the instructional role, nor may they match the learning style of the student.
The notion of a Learning Environment as Wilson (Wilson 1995) notes, at minimum suggests that we are considering the learner being positioned in, and engaging with, some setting where they can interact via devices and tools, collecting and interpreting information, interacting perhaps with others in principally a self directed manner. The subsequent focus in this LE is on the learner with learning being fostered and supported.
Unlike other settings supported by technology eg Computer Assisted Instruction and Intelligent Tutoring Systems to a degree, LE’s are not pre-packaged and prescribed, and a “level of uncertainty and uncontrolledness comes into play” (Wilson 1995). It is the role of the teacher then, to create or locate instructional artefacts and maintain the environment so that successful learning can take place.
Some desirable characteristics of emerging LEs as suggested by Maurer (Maurer 2002) include:
· re-usable content modules accompanied with searchable metadata;
· facilities that allow the use of the system for various learning paradigms and different levels of learners,
· a range of tools to create, combine and modify such modules into new ones;
· tools that allow the administration of the modules mentioned with statistical data gathered for authors, teachers, tutors and students as the system is used;
· a set of features for communication and collaboration including chats and discussion forums;
· provision of a high degree of interactivity with powerful tools for (self) testing and feedback.
A learning environment then, is a domain populated with instructional items, presented as either multimedia of hypermedia objects with some addressing mechanism and supported by a group of tools. The address might be a URL, or a directory path to a locally stored object. The term object is used here to conform to the IEEE Learning Technology Standards Committee specification (Committee 2000) for learning objects and there is general agreement with (Wiley 2001) interpretation as ‘any digital resource that can be reused to support learning’.
Adaptive Hypermedia (AH) with its strong links to the Intelligent Tutoring Systems (ITS), utilise fine grained hypermedia educational modules with associative links to others. The further coupling of student and course profiles, has resulted in AH learning environments that provide the foundations of student centred work places. A number of universities are pursuing this development path in an attempt to provide managed, personalised learning environments for students.
These systems include a semantic network of concepts to represent the domain knowledge (domain model); some AH systems [ISIS-Tutor, Hyperadapter, ELM-ART, Anatom-Tutor) (Brusilovsky 2000) also use frame-like representations with sets of attributes to provide a finer resolution of a concept. The ability to present different content through the variable use of attributes, constitutes adaptive presentation.
A concept can also consist of a combination of documents (Pilar da Silvar, Van Durn et al. 1998). JointZoneã (Ng 2002) utilises both declarative and procedural knowledge in a hypermedia learning that provides adaptive case studies to medical students. Mullier inserted tutorials into the semantic network making up Hypernet (Mullier and Dixon 2000) to provide the students with tutorial tasks to reinforce learning.
Typed links between these concepts are used to define relationships. The kinds of links of typical systems (HypaAdapter, EPIAIM, Anatom-Tutor, KN-AHS, ELM-ART) (Brusilovsky 2000) include is-a, pre-requisite and part-of. Adapting at the link level is refered to Adaptive Navigation Support (ANS). This is a generic term used for a set of techniques used in AHS to provide directional assistance to a user by altering the appearance of observable links (Brusilovsky 1996). The main adaptive navigation techniques in use today include direct guidance (next, sorting, link hiding) and annotation (Pilar da Silvar, Van Durn et al. 1998). Annotation typically employ some metaphor eg traffic light metaphor to indicate a users readiness for a concept.
The domain model provides a structure for the representation of the user’s domain knowledge. For each concept within the domain, a value is maintained to indicate/estimate the users level of knowledge of the concept. This value could be a boolean, a qualitative value, or some quantitative estimation of the probability that the user knows the concept. In total, it constitutes an overlay model for the user in the domain that drives both adaptive presentation and navigation.
Stereotyping (Kay 2000) employed within the student profile, provides a mechanism for the administration of content that varies in complexity. Student access to these settings then provides an adapatable (Oppermann 1997) environment for students who wish to scrutinise and modify their settings.
The concepts of the AHS described above constitute learning objects as defined by IEEE Learning Technology Standards Committee specification; however, there has been a growing concern among instructional designers that the metadata specification for the instructional significance of a learning object within the IEEE specification is inadequate. Wiley (Wiley 2001) proposed a taxonomy for learning objects that was generic in the sense that it did not specify the role of the object in the instructional process.
Jonassen et al (Jonassen, Peck et al. 1999) take the position that the true worth of multimedia and hypermedia might be obtained through the learner constructing knowledge via the use of technology, rather than as a mode of delivery. Although there is agreement that the creation of multimedia and hypermedia artefacts is a powerful mechanism for individual learning, the learning benefits attributed to the construction of different media views of content cannot be discounted; dual coding theory supports this assertion and recent experiments (Alty 2002) indicate the significance of media in the presentation process.
Beacham et al. (Beacham 2002) further investigated the correlation between individual learning styles as defined by Felder (Felder 1993) with respect to the student use of multimedia presentation and learning materials within a learning environment. An index of learning styles questionnaire was used in this case to map students to Felder’s taxonomy. Others (Carver 1996; Wolf 2003) have further mapped learning styles into there adaptive engines.
It has been suggested (Bannan-Ritland, Dabbagh et al. 2001) that a Cognitive Flexibility Hypertext (CFT) (Spiro 1987) with multiple knowledge representations, is a good match for a hypermedia learning environments populated with learning objects. “This method of representing instructional content through intersecting themes and cases spawns the constructivist principle of knowledge construction requiring learners to assemble a flexible schema that is situation specific” (Spiro 1987).
This theory is underpinned by constructivist (Jonassen 1997) principles with the students actively involving themselves in the learning process, creating items and placing them into their workspace or on the WWW, searching out content and skills, and satisfying authentic tasks. These educational tasks in total, provide the macro level scaffolding (Bannan-Ritland, Dabbagh et al. 2001) that allow the students to use particular content in different, albeit overlapping, contexts.
Other characteristics of CFT (Bannan-Ritland, Dabbagh et al. 2001) include:
· explicitly linking and tailoring concepts to practice and case examples (i.e. situating conceptual knowledge in contexts that are similar to those required for the application of the knowledge);
· incrementally introducing complexity in small, cognitively manageable units;
· stressing the interrelatedness and web-like nature of knowledge (instead of isolated and compartmentalized knowledge); and
· encouraging the assembly of appropriate knowledge from various conceptual and case resources (rather than the intact retrieval of previously memorized information).
CFH then is a conceptual model (Jonassen 1997) for designing learning environments that is complementary to other learning theories, in particular those grounded in constructivist principles and based upon cognitive learning theory. Some characteristics of learning environments derived from this conceptual framework include (Jonassen 2003)
· an appropriate degree of complexity faced by practitioners;
· multiple representations of content;
· emphasises case-based instruction;
· context dependent knowledge;
· knowledge construction, not transmission.
The majority of these characteristics are espoused in other constructivist based theories of instruction (anchored instruction, cognitive apprenticeship, situated learning) while not excluding others like dual coding theory (Alty 2002) for example.
Allert et al. (Allert 2002) have proposed a top down framework with four levels of abstraction to define the pedagogical dimensions of each learning object. Each learning object may play different roles in the learning process, depending on the learning theory and the instructional design that it is used within. They suggest that instructionally oriented classifications should be embedded in the Learning Object Metadata (LOM) (Committee 2000) to facilitate use by both teachers and prospective students. Allert et al. further use expository teaching and problem based learning as exemplars and examine the phases within each instructional model.
This framework then, has implications for instructors in that it will catagorise learning objects, learning theory and instructional purpose irrespective of the media form that they might take. The relationship between learning theory and instructional purpose need not be a one to one relationship; infact, one instructional piece may have different roles in different paradigms. This meta data will be able to be searched for by both instructors and learners and the object subsequently reused.
To support other teaching methodologies, the metadata defining instructional objects needs to be searchable; tools like search engines, glossaries and concept maps have been employed (Brusilovsky, Eklund et al. 1998; Kurzel 2002) to allow the students to locate content. With metadata concerning the instructional role, tools might filter returned information depending on the teachers, or indeed the student’s, instructional preference.
The students need to also have access to all of the information about both their interaction with the course content and their progress through it. Tutors too need convenient access to the central repository of both course and student information. Reporting mechanisms need to be in place for both students and instructors to display relevant information. Educational hypermedia systems typically account for this through the use of the course, student, and assessment models.
Communications facilities should also exit that for example, allow on-line tutorial groups to discuss problems and assignments, or to simply discuss particular content-based issues. Tutors should also have the ability to address any concerns that are not accounted for by peer group interactions; the use of a real time synchronous electronic helpdesk interacting one on one, or indeed with some broadcasting functionality is seen as advantageous.
AMLE, the LE under consideration, is made of a number of components that in total, constitutes a social system that needs to be considered as a whole. The human side of the system includes the instructors and the students, while the learning objects, adaptive engine, display systems, search tools, authoring tool, metadata etc., constitute the networked information system.
Fig 1: AMLE Architecture
The students interact with the educational domain through a hybrid browser that supports:
· fine grained content presented as concepts (Murray, Condit et al. 1998);
· typed links between concepts (pre-requisite, is-a-practical, etc.);
· a range of document and media formats supporting content (Pilar da Silvar, Van Durn et al. 1998);
· course and student overlay model driving adaptive annotations;
· competency based assessment employed at the concept level;
· a range of tools allowing the searching of content (Brusilovsky 2000);
· the dynamic WWW delivery of instructional templates/assignments;
· capture of episodic data defining the student traversals of the domain;
· facilities for student/student and student/instructor interactions.
Fig 2: Session Viewer
The domain knowledge is initially administered in weekly sessions via the Session Viewer which constitutes the expert view.
Fig 3: Searching and bookmarking
Each concept can then be individually viewed through a Concept Viewer. Search tools exist to allow the instructor and student to break away from this expert view and allow for more student centred instructional methodologies. The instructional model and subsequent artefacts are delivered on-line as required; these include formative tasks/tests and summative assignments
The prototype of the main interface (AMUI in figure 1 above) has been implemented in Macromedia Director (the main development environment used in the course), so that students can add/modify the interface as required. Multimedia artefacts can be included in their environment, enabling some personalisation at the interface level. The resulting assignment/project work can also be embedded into their interface.
The content domain currently accounted for, is the content associated with Multimedia Concepts, an introductory course within the multimedia studies stream offered at the University of South Australia. The aim of this course is to provide the foundational knowledge and skills required to create and utilise a range of media items within multimedia presentations.
Course materials are online but are not fine-grained enough to enable their reformatting to take account of any existing knowledge and learning style. The students undertaking the course come with a diverse range of knowledge and skills. The course is situated with an Arts degree, but a large number of students from Computing, Education, Art/Design to name a few, undertake it. These students can also come from different levels/years in their study programme.
Information Technology based domains are ideally suited for constructivist LEs because authentic activities are generally emphasised. Students are placed in real-world situations and need to acquire the knowledge and skills required by practitioners. These can involve:
· the representation and construction of media artefacts;
· problem solving methodologies;
· project development methods;
· design representations;
· graphical design skills;
· information structuring;
· programming concepts;
· communication of information;
· consideration of ethical issues.
An authentic activity might require any number of the above in their design and creation.
The adaptivity employed in educational hypermedia systems via a student overlay model provides the mechanism for personalisation at the content level. AMLE provides annotated associative links within the system that can either address the WWW in total, or the local file space. Media rich items can be stored locally to avoid problems associated with delay. Link hiding (Brusilovsky 1996) is not employed; instead, annotations provide relevant information to allow the user to decide whether they are to follow the link.
Stereotyping (Kay 2000) within the student modelling structure, appears to be an appropriate technique where distinct student groups with similar characteristics can be identified. The coupling of different content presentations to the stereotyping system allows for variance in the content presented (adaptive presentation (Brusilovsky 1996)). An experienced computer science student when dealing with a conditional construct for example, may simply wish to see a brief explanation of the syntax. Novice programmers would need more explanation with an example perhaps. However, any system based on scrutable stereotypes and preferences, should be observable and thus adaptable (Oppermann 1997)
Assessment at the concept level is based on a competency model which is administered online; where a concept is a skill or task, students are asked to complete a task and have it registered manually. For example, the concept we are dealing with might be the scanning of an image. The concept might be demonstrated and the student asked to perform the task; completion of this task is manually registered. Scrutability should also pertain to both the course model (pre-requisites, outcomes) and the formal assessment, and students should be able to view their standing in the course at anytime.
Pre-course testing can also be used to initiate the compentency model. On-line testing can map into the competency model with some inferences being made concerning knowledge. A range of practical skills can be accounted for manually through the demonstration of practical artefacts constructed by the student. The classification of learning style would best be made pre-course through access of Learning Styles Questionnaires.
The goal is to create instructional objects that allow the instructor to use the hypertext/hypermedia learning environment and employ other instructional methodologies. Wiley (Wiley, 2002) would define the session viewer as a generative object that organises concepts into expert specified groupings. The finer the resolution of each of the objects, the more likely it is that each concept is simply a declarative statement about a concept, or a procedural specification with the emphasis being on its practical use.
Further, we can create learning object templates that make use of the multimedia/hypermedia objects and consequently, represent phases within a specific learning theory. For example, the phase background knowledge (Allert 2002) within the Problem Based Learning (PBL) theory, requires the identification and location of relevant information. The template in this case might provide information about what should be searched for and how to go about it in the learning environment. Another object might describe the algorithm involved.
Within AMLE, an instructor might decide to deliver the course using PBL; instead of providing a week by week prescription of the declarative and procedural concepts with weekly tasks to perform, students might be presented with an ordered sequence of learning objects that provide information about each of the phases. Each phase might have some formative assessment component to be satisfied.
The fact that a learning objects can have a role within a number of theories, constitutes a network. The fact that some phases precede others is analogous to the pre-requisite relationship of the concepts within AMLE. With more instructional methodologies defined in terms of phases with corresponding learning objects in the LE, the student too might select a particular methodology; here the LE would automate the delivery by searching the network of objects.
Our major concern has been to provide adaptivity and interactivity for our students since we recognise that our students’ backgrounds and cultures have given us a wide range of learning styles and expectations among our students. Currently, we have created a range of hybrid browsers with varying interfaces to access the course model; a number of players have been created to display the content and associated materials.
We are continuing to investigate the following tenets (Jonassen 2003) of constructivist learning environments within AMLE:
· how learners can be actively engaged in the learning process;
· tools and activities that promote the construction of new student knowledge in student centred ways;
· how learners can work collaboratively;
· ensuring engagement in complex and ill-structured contextual problems;
· how the LE best supports the conversational process within the LE;
· how best to provide for student reflection of learning;
with the aim being the development of a LE that provides multiple views of the content domain.
Other instructional methodologies, and educational activities underpinned by the constructivist philosophy (in particular problem solving) are being explored. So too are the different tools that will be available within the interface. Both learning style and stereotyping data which have the potential of feeding back into the system and enhancing the adaptivity employed, are being investigated.
We have established a hypermedia/multimedia learning environment that adapts to prior knowledge through the use of link annotation; we have also introduced stereotypes into the overlay student model and have coupled this to different content presentations. Alternate instructional methodologies we think can be embedded within the LE as hypermedia artefacts, each corresponding to a stage within the methodology. Similar tools and metadata employed with the concept-based AH environment can be employed to utilise these. A range of other tools like search engines, indexes, help, tutorial groupings etc. will support these other instructional methodologies.
The dynamic components of the learning environment i.e. assessment, course structure, instructional methodology etc. are stored on a remote server and are accessed through the WWW. Tutors have convenient on-line access to student groupings to enable marks entry. Student preferences and settings, along with marks etc that make up the assessment profile of the students model, are available to the student.
Allert, H., Dhraief, H, Nejdl, W. (2002). Meta-Level Category 'Role' in metadata Standards for learning: Instructional Roles and Instructional Qualities of Learning Objects.
Alty, J. L. (2002). Dual Coding theory and Computer Education: Some Media Experiments to Examine the Effects of Different Media on Learning. ED-MEDIA, Dever, Colorado, AACE.
Bannan-Ritland, B., N. Dabbagh, et al. (2001). Learning Object Systems as Constructivist Learning Environments: Related Assumptions, Theories and Applications. The Instructional Use of Learning Objects. D. A. Wiley. Bloomington, IN, Association for Educational Communications and Technology.
Beacham, N. A., Elliott, J.L., Alty, J.L., Al-Sharrah, A. (2002). Media Combinations and Learning Styles: Dual Coding Approach. ED-MEDIA, Denver, Colorada.
Bonkhorst, J. (1998). The use of multimedia learning environments in teacher training colleges. ATEE, Limerick, Ireland.
Brusilovsky, P. (1996). "Methods and Techniques of Adaptive Hypermedia." User Modeling and User-Adapted Interaction Vol 6(2-3): 87-129.
Brusilovsky, P. (2000). Adaptive Hypermedia: From Intelligent Tutoring Systems to Web-Based Education. Intelligent Tutoring Systems, Montreal, Canada, Springer-Verlag.
Brusilovsky, P., J. Eklund, et al. (1998). "Web-based Education for all: a tool for development adaptive courseware." Computer Networks and ISDN Systems 30: 291-300.
Brusilovsky, P., E. Schwarz, et al. (1996). A Tool for developing adaptive electronic textbooks on the WWW. Proceedings of WebNet'96, World Conference of the Web Society, San Francisco, CA.
Carver, C. A., Richard, A. H. & Edward, L. (1996). Enhancing Student learning by Incorporating Learning Styles into Adaptive Hypermedia. ED-MEDIA96, Charlottesville, Virginia, AACE.
Committee, I. L. (2000). Learning Objects Metadata Draft v4.1.
Eklund, J., and Brusilovsky. P. (1998). Individualising Interaction in Web-based Instructional Systems in Higher Education. AUC Academic Conference, University of Melbourne, Australia.
Felder, R. (1993). "Reaching the Second Tier: Learning and teaching in College Science Education." J. College Science Teaching, 23(5): 286-290.
Gibson, S. (2002). "Using a problem based multimedia enhanced approach in learning about teaching." Australian journal of Educational Technology 18(3): 395-409.
Herrington, J. and R. Oliver (1999). "Using situated learning and multimedia to investigate higher-order thinking." Journal of Interactive Learning Research 10(1): 3-24.
Jonassen, D. (2003). Welcome to the Design of Constructivist Learning Environments. 2003.
Jonassen, D., K. L. Peck, et al. (1999). Learning with Technology: A Constructivist Perspective. Upper Saddle, NJ, Merrill, Prentice Hall.
Jonassen, D. H., Dyer, D., Peters, K., Robinson, T., Harvey, D., King, M., and Loughner, P. (1997). Cognitive Flexibility Hypertexts on the Web: Engaging Learners in Meaning Making. Web-based Instruction. B. Khan. Englewood Cliffs, NJ, Educational Technology Publications: 119-134.
Kay, J. (2000). Stereotypes, Student Models and Scrutability. ITS2000, Montreal, Canada, Springer-Verlag.
Kurzel, F., Slay, J. & Chau, Y. H. (2002). Towards An Adaptive Multimedia Learning Environment: Enhancing the Student Experience. ED-MEDIA2002, Denver, Colorada, AACE.
Looi, C. K. and D. Ang (2000). "A multimedia-enhanced collaborative learning environment." Jouarnal of Computer Assisted Learning 16: 2-13.
Maurer, H. (2002). What Have we Learnt in 15 years About Educational Multimedia. EDMEDIA 2002, Denver, Colorado, AACE.
Mullier, J. L. and M. B. Dixon (2000). Authoring Educational hypermedia Using a Semantic Network. ED-MEDIA 2000, Montreal, Canada, AACE.
Murray, T., C. Condit, et al. (1998). A Preliminary Framework for Concept-Based Adaptive Hypermedia. ITS-98 workshop on WWW-Based Tutoring.
Ng, M. H., Maier, P. and Hall, W. (2002). Makiung Web-based Learning Adaptive. InSITE2002, Cork, Ireland, AACE.
Oppermann, R., Rashev, R. & Kinshuk (1997). Adaptability and Adaptivity in Learning Systems. Knowledge Transfer. A. Behrooz. London, UK, pAce. 11: 173-179.
Pilar da Silvar, D., R. Van Durn, et al. (1998). Concepts and documents for adaptive educational hypermedia: a model and a prototype. 2nd Workshop on Adaptive Hypertext and Hypermedia, HYPERTEXT'98, Pittsburgh, USA.
Ricketts, J., F. H. Wolfe, et al. (2000). "Multi-media - Asynchronous distributed education." Social science computer review 18(2): 132-146.
Spiro, R. J., Feltovitch, P, J., Jacobson, M.J., & Coulson, R.L. (1987). "Cognitive Flexibility, Constructivism, and Hypertext: Random Access Instruction for Advanced Knowledge Acquisition in Ill-Structured Domains." Educational Technology: 24-33.
Wiley, D. A., II (2001). Connecting learning objects to instructional design theory: A definition, a metaphor, and a taxaonomy. The Instructional Use of Learning Objects. D. A. Wiley. Bloomington, IN, Association for Educational Communications and Technol ogy.
Wiley, D. A., II (2002). C. Learning objects need instructional design theory, IN, The 2001/2002 ASTD Distance Learning Yearbook, McGraw-Hill, New York.
Wilson, B. G. (1995). "Metaphors For Instruction: Why We Talk About Learning Environments." Educational Technology 35(5): 25-50.
Wolf, C. (2003). iWeaver: Towards 'Learning Style'-based e-Learning in Computer Science Education. ACE2003, Adelaide, South Australia, Australia, Conferences in Research and Practice in Information Technology,.