Authoring for Adaptive Presentation Authors: Adam Moore, University of Nottingham, University Park, Nottingham, Notts., NG7 2RD
[email protected] Tim Brailsford, University of Nottingham, Jubilee Campus, Nottingham, Notts., NG8 1BB
[email protected] Craig Stewart, University of Nottingham, Jubilee Campus, Nottingham, Notts., NG8 1BB
[email protected] Peter Davies, University of Nottingham, Jubilee Campus, Nottingham, Notts., NG8 1BB
[email protected]
Abstract An increasing number of adaptive hypertext systems are being used in education, for example, to deliver courses appropriately for each learner, or to deliver a tailored support experience. However, there is little in the way of guidelines available to content authors, on how to best tailor their material for delivery in such systems. This paper will attempt to address this lack by outlining some simple approaches to the preparation of such material. Firstly, we will examine sourcing material, especially the conversion of extant, legacy sources. Next we will expound on the many assumptions made when writing non-adaptive material, and attempt to show where these are also appropriate in writing adaptive material, and where their application is counter productive. Finally, we will address specific authoring requirements that are required to deliver material suited for the different types of adaptation that takes place in modern systems.
Index Terms Adaptive learning, authoring guidelines, content authoring, virtual learning environments INTRODUCTION An increasing number of adaptive hypertext systems are being used in education, for example, to deliver courses appropriately for each learner, or to deliver a tailored support experience. However, there is little in the way of guidelines available to content authors, on how to best tailor their material for delivery in such systems. The following paper will attempt to address this lack by outlining some simple approaches to the preparation of such material. It should be noted that this is general advice, and may need to be further tailored to individual applications. This document is not a discussion of the pros and cons of adopting electronic delivery of learning material, nor does it discuss the balance between electronic and face-to-face teaching, both of which are important issues, which must be considered by educators wishing to utilise this technology. It is, however, based on real experience, preparing material for WHURLE, the Web-based Hierarchical Universal Reactive Learning Environment [1-3] and research currently underway for the ADAPT project [4]. Types and Methods of Adaptation There are more and more learning environments with some level of adaptation becoming available, but they can be simply divided by the appraoches they take to delivering an adaptive experince: • Link / Navigational adaptation – Here the adaptation is in the path that the user is allowed to take/view through the material, where a link to more advanced material may be hidden or deprioritised (e.g. moved further down a page or menu) until it is appropriate to present it to the learner. • Structural adaptation – Here the adaptation is concerned with altering the structure of the material the learner encounters. • Content adaptation – Here each learner is presented with the appropriate content. Content that is not appropraite is hidden. As the learner progresses, new material may become available and old material inaccessible. • Presentational adaptation – This type of adaptation is usually done on the basis of either particular learning styles of the learner, or to address particular interface issues, such as accessibility or localisation (e.g. menus in the appropraite language). Models of Learners and Authoring Strategies Adaptation is based on a system responding to the inbuilt model of the user of the system (learner). There are currently two most commonly used types of user model: Stereotyping – here each learner is assigned to a group, each group having particular properties. International Conference on Engineering Education
July 21–25, 2003, Valencia, Spain. 1
Overlay model – here the knowledge of the learner in each domain is used to determine what material is appropriate Most systems will use one or the other, and it is important to know which model is being used in your target system, so that your material takes this into account. An interesting system which uses a combination of both is the hybrid model, developed by Zakaria et al at the University of Nottingham [5,6]. Consciously or subliminally, learner models shape the way all teachers do their work. In some contexts, these models are explicit and consistent with the findings of research, but more often (certainly in higher education) they are tacit, unsupported by evidence and rooted in traditional disciplinary practices that seem impervious to criticism and resistant to change [7,8]. Most of the computer-based learning developments in UKHE during the 1990s suffered from the same uncritical acceptance of outdated traditional approaches to teaching and learning (the ‘information transmission’ model). Against this background, what are the lessons for the next generation of technology-mediated learning support systems, those that seek to accommodate learner diversity through the use of adaptive technologies? To be effective, an adaptive learning system (ALS) must know the learner, know about learning and know how to respond when the learner is engaged in learning. Knowing the learner means having access to relevant personal information about the learner’s background, capabilities and needs (the ‘user profile’). Knowing about learning means using a learner model (a) that embodies what is known about learner diversity and the psychology of learning (learning theories [9], cognitive styles, learning styles and learning strategies [10]) and (b) that accommodates creatively what is not yet known about the workings of the learning mind (a great deal). Knowing how to respond means being able to match system output to user need Since learning theories abound and learning styles are diverse and our ignorance is profound, the challenge confronting ALS designers may seem insurmountable. Furthermore, adaptive learning systems require adaptable content, and as yet there are no discipline-independent, pedagogically credible rules or systems for producing it. A core requirement in this context is a credible developmental model of learning: is it like a string of beads that grows longer as new knowledge is acquired, or more like a living organism that grows and become internally more complex as it develops? It matters, because each model has different implications for the authoring process. Could it be that both models are applicable in different contexts, in which case authoring becomes more complex still. We should pause for a moment to consider what a working ALS should be, in metaphorical terms. Should it be a surrogate lecturer or a responsive and adaptable learning aid? A ‘sage on the stage’ or a ‘guide on the side’? The distinction matters here too, because the former would be more costly to implement and much less effective. A ‘surrogate lecturer’ would embody a familiar but outdated pedagogy according to which teaching is a process of transmitting information / knowledge / wisdom, and learning is a process of passive absorption and assimilation. Content volumes and production costs in such systems tend to be high, while learning effectiveness tends to be low. Much of the courseware produced in UK universities during the nineties was like this. Research, however, points in the opposite direction: that teaching is most productive when it is learner-centred; and that learning is most profound when it is active, independent, enquiry-based and critical. The implications of this insight for ALS authors are clear: powerful though modern technologies may be, learning systems are tools, not teachers, and their effectiveness will depend on the extent to which they allow the learner to remain in control. That being the case, the shortcomings of our explicit understanding of how deep learning occurs will be offset by the learners’ tacit (intuitive) understanding of how best to get to grips personally with new challenges and learning opportunities. The role of the ALS in this model is therefore to facilitate learning, not by teaching didactically but by providing the ‘personalised’ scaffolding within which deep and independent learning can occur. Such a scaffolding would, among other things, motivate the learner and respond to ongoing needs, and it would be generated, not by the author alone but by an authoring system he has been trained to operate productively. Against this background, the authoring challenge becomes a little less daunting. Subject specialists who become authors would not be expected to become pedagogues or experts in adaptive technology. Neither would they be expected to produce large volumes of didactic text. Rather they will be required, under system guidance, to specify objectives, delineate conceptual frameworks and learning pathways, identify problem-solving opportunities and relevant online and other resources, produce commentaries and guidelines for the proper use of these resources, and devise strategies for assessment. Much of this contribution would be delivered via a ‘knowledge-eliciting’ form designed in accordance with accepted pedagogical principles but not necessarily understood in every detail by the content provider, at least initially. In most cases, some training would probably be desirable and more efficient. But crucially, the authoring system would ultimately become part of the learning environment itself, one of the many creative tools at the disposal of both author and learner. In this more learner-centred system, the burden of responsibility for gathering, delivering, interpreting and ultimately assimilating information would, to a significant extent, shift from the teacher / author to the active, semi-independent learner. Underpinning the learning system’s adaptive responses would be a structuring and guidance technology responding to the combined effect of (i) the learner’s user profile, (ii) the course’s learning objectives, and (iii) the user’s decision-making International Conference on Engineering Education
July 21–25, 2003, Valencia, Spain. 2
activities during engagement with the system, which would generate adaptation and feedback. Nebel et al list and compare some sixteen adaptive methods and algorithms, the most common two of which (Stereotypes and Overlay Model) are described above [11]. They classify the methods into three broad categories: stereotype-based, rule-based and mathematical & statistical approaches. How generic and transferable these technologies are is not clear. Weber describes an adaptive learning system developed for use on the web, and reports some positive early evaluations [12].
SOURCING MATERIAL The first consideration when preparing material for delivery via an adaptive learning environment is where to source such material from. The next sections cover the two options available – converting extant legacy material, or starting from scratch. Legacy Material There has been an enormous growth in the quantity of material prepared for electronic delivery in the last fifteen or so years, propolled by an increased interest in both educators and administrators in utilisng this delivery format, and by government initiatives, such as TLTP (Teaching and Learning Technology Programme [13]) in the UK. However, in many cases this material has been created and locked into proprietary, closed formats, making reusability of this material a difficult process. Here are some simple guidelines to ease the conversion: • Don’t throw anything away – although your current target of the conversion may not require some of the information, structure or meta-information in your legacy material, you may have a use for it in the future. • Use a long-term intermediate – whilst the target of your conversion may be to one of the closed, proprietary platforms, in order to gain maximum benefit from your conversion process, it is important to have your material in a form which allows you to reuse it again, if your delivery platform subsequently changes, or if you wish to distribute your material. In the authors’ opinion, the current most suitable formats for this are XML [14] and SQL-compliant databases (such as MySQL [15]) which allow both structure and content to be represented. • Use standards wherever possible – it is important that your material conforms to the current standards of both your domains (e.g. the appropriate ANSI [16], BSI [17], CAM-I [18],etc. protocols) and the format into which it is to be converted. If your intermediate format is XML or a database, try and find the appropriate schema and STICK TO IT. • Convert or generate as much meta-data as you can – the information about your material is a valuable commodity – if it is already present in your system, preserve it, if not, take this opportunity to generate it. Also, use the appropriate standard for your metadata – utilising IEEE Learning Object Metadata standard [19] and the IMS Learning Resource Meta-data specification [20], or the appropraite geographical instantiation of these protocols, such as CanCore [21] in Canada. Creating New Content Coming to create new content for an adaptive learning environment from scratch has both advantages and disadvantages. Of course, the main advantage is the avoidance of the problems associated with the conversion of legacy material, as described above. Also, the conversion from legacy material will inevitably not produce content that can be readily adapted (due to the educational nature of the source material). However when creating from scratch this is an important consideration. The disadvantage is, of course, that it will take much longer – a good estimate is about 20 hours of preparation time for each instructional hour of electronic teaching material. This may be much longer if multimedia objects are being prepared. It should also be remembered that when preparing adaptive material, an hour of learning material must be prepared FOR EACH LEARNER MODEL – although it is probable that much material can be reused, this important additional time burden must be considered. • •
Appropriateness to curriculum – many people see using adaptive learning environments as an excuse to include all material they can find on the subject in question into a course. This is often a time-consuming, open-ending task. Stick to the appropriate curriculum or lesson plan. Appropriateness to audience – it is important to remember that there is still a specific group of people that the material is aimed at – ensure their needs are considered during the development of the material.
AUTHORING FOR VIRTUAL LEARNING ENVIRONMENTS Teachers who have much experience in the design of classroom materials have valuable skills in the general design and presentation of educational material, but sometimes struggle when confronted with a totally new presentational format. The key thing to remember about authoring for electronic delivery is that the material must be able to stand on its own, that is, it
International Conference on Engineering Education
July 21–25, 2003, Valencia, Spain. 3
must make sense without any further explanation. Many people coming to author electronic material for the first time make the mistake of assuming that there will always be an opportunity for elucidation for any material they produce. Accessibility Whilst studies have shown that using electronic material can cause a gap in attainment, as some students are better suited to this method of learning [22], it is important to remember that some students have accessability issues with any electronic medium [23]. In many countries, there is specific legislation to ensure equal access to all students to all materials (e.g. SENDA [24] in the UK). As well as obtaining general help from groups such as the W3C’s, web accessability initiative [25], There are also usually bodies set up to help educators address the requirements of their particular country’s legislation (e.g. TechDis [26] in the UK)
AUTHORING FOR ADAPTIVE PRESENTATION The exact authoring for your material will depend on the system you are preparing material for, and so will not be covered here. Instead, what follows is advice on preparing material suitable to be adapted for presentation in adaptive environments in general. We assume that (at least) a two stage process is used for authoring – firstly, subject experts prepare content, then teachers prepare lessons. Content Authoring The key to authoring content for adaptive systems is to ensure that the resulting material is suitable for the method of adaptation the target system employs. In most cases, this will entail writing short, discrete ‘chunks’ of information. This process of ‘discretisation’ is often difficult for the new author, especially one immured to the construction of long, discursive prose. Each chunk must be able to stand on its own, as the adaptive system may, or may not, include other chunks at any one time, so it is important that the material is self-contained. This means that there must be no references such as ‘see further on’, or ‘this will be discussed later’ and such like within the chunk. It is helpful to think of the process as one of creating ‘atomic units of information’, that is, the smallest, self-contained piece of knowledge that can reasonably be constructed. Examples of such information include a paragraph or two of text, or perhaps a captioned image. Of course, chunks do not have to be small in size; an image on its own is usually meaningless, unless it contains explanatory text, and so a commentary or descriptive caption is almost always an essential addition to any chunk containing multimedia. Also, discretisation must be performed in a sensible manner; a legal document, which contains much internal cross-referencing and reflections may be a very large document, but only a single chunk. Similarly, the derivation of a mathematical formula. It is also important to realise that each chunk can be written in whatever style the author wishes – discursive, didactic, elaborate, etc. But it must be realised that the mixing of styles within a lesson may cause comprehension difficulties with your learners. Many authors find that a simple, straightforward relation of facts to be an effective writing style, with occasional ‘personal’ reflective chunks available to allow the insertion of a discursive style of lesson, so that the learner is not simply guided through a maze of facts. Lesson Authoring When constructing adaptive lessons from available material, it is important to utilise the adaptive models available in an efficient and effective manner. Think about the different models of students that you are providing material for and ensure that the paths through the material you construct are suitable, but also provide sufficiently similar material to allow any evaluation or assessment that you perform to be carried out in an equitable manner. Although little research has been performed in providing software support for adaptive authoring, the REDEEM system supports this approach [27-29]. REDEEM depends on human teaching tactics and strategies. Its design is predicated on the view that teachers should be provided with the means to express their theories about how learners should be taught and that ITSs should then teach students according to these theories.
CONCLUSIONS This paper has described some simple ways in which authors can prepare material for inclusion in adaptive electronic learning environments. We hope this document proves useful and welcome all feedback on its appropriateness. We also welcome any additions that interested parties would like to make, either for specific systems, or for sections that this paper has not covered in depth.
International Conference on Engineering Education
July 21–25, 2003, Valencia, Spain. 4
ACKNOWLEDGEMENT The authors would like to thank the Nottingham Web Technology Group for their many helpful contributions to this document and related works. Craig Stewart is funded by the ADAPT project, which is carried out with the support of the European Community in the framework of the Socrates programme.
REFERENCES [1]
Moore, A., Brailsford, T.J. Stewart, C.D. “Personally tailored teaching in WHURLE using conditional transclusion”, Proceedings of the Twelfth ACM Conference on Hypertext and Hypermedia (Hypertext ‘01). 2001, pp 163-164.
[2]
Brailsford, T.J.; Stewart, C.D.; Zakaria, M.R. Moore, A. “Autonavigation, Links and Narrative in an Adaptive Web-Based Integrated Learning Environment”, Eleventh International World Wide Web Conference (WWW2002), Honolulu, Hawaii, 7-11 May 2002. Available via: http://www2002.org/CDROM/alternate/738/.
[3]
WHURLE Homepage – http://whurle.sf.net
[4]
Minerva project: ADAPT (project reference number 101144-CP-1-2002-NL-MINERVA-MPP*) "Adaptivity and adaptability in ODL based on ICT"
[5]
Zakaria, M.R.; Moore, A.; Ashman, H.; Stewart, C.D. & Brailsford, T.J. “The Hybrid Model for Adaptive Educational Hypermedia” Second International Conference on Adaptive Hypermedia and Adaptive Web Based Systems, May 29-31, 2002, Malaga.
[6]
Zakaria, M.R. & Brailsford, T.J. “User Modelling and Adaptive Educational Hypermedia Frameworks for Education”, New Review of Hypermedia and Multimedia (in press).
[7]
Gibbs, G “Twenty Terrible Reasons for Lecturing” SCEDSIP Occasional Paper No. 8, 1982
[8]
Ramsden, P “Learning to Teach in Higher Education” Routledge, 1992
[9]
Kearsley, G “Explorations in Learning & Instruction: the Theory into Practice Database”, 1994-2003 available from: http://tip.psychology.org/index.html
[10] Riding, R and Rayner, S “Cognitive Styles and Learning Strategies: Understanding Style Differences in Learning and Behaviour” David Fulton, London, 1998 [11] Nebel, IT et al “User ontologies for adaptive interactive software system”, 2003 available from: http://ontology.buffalo.edu/medo/Adaptivity.pdf [12] Weber, G “Adaptive learning systems in the World Wide Web”, 1997 available from: http://www.cs.usask.ca/UM99/Proc/invited/Weber.pdf [13] TLTP Homepage - http://www.ncteam.ac.uk/projects/tltp/ [14] Extensible Markup Language (XML) 1.0 (Second Edition) W3C Recommendation 6 October 2000, Tim Bray, Jean Paoli, C. M. Sperberg-McQueen, Eve Maler [15] MySQL Homepage - http://www.mysql.com/ [16] American National Standards Institute - http://www.ansi.org/ [17] British Standards Online - http://www.bsi-global.com/Ways+to+Purchase/BSOL/index.xalter [18] Consortium for Advanced Manufacturing - International - http://cam-i.org/standards.html [19] IEEE P1484.12.2/D1, 2002-09-13 Draft Standard for Learning Technology - Learning Object Metadata - ISO/IEC 11404 [20] IMS Global Learning Consortium: Specification: Meta-data - http://www.imsglobal.org/metadata/index.cfm [21] CanCore - http://www.cancore.ca/ [22]
Quentin-Baxter, M, “Quantitative evidence for differences between learners making use of passive hypermedia learning environments” ACM Computing Surveys, 31, 4., 1999. Available via http://www.cs.brown.edu/memex/ACM_HypertextTestbed/papers/52.html
[23] Nielsen, J, “Beyond Accessibility – treating users with disabilities as equals”, Alertbox, Nov 11, 2001, available from http://www.useit.com [24] Special Educational Needs and Disability Act 2001, ISBN 0 10 541001 2 – available from http://www.hmso.gov.uk/acts/acts2001/20010010.htm [25] W3C Web Accessibility Initiative - http://www.w3.org/WAI [26] TechDis - http://www.techdis.ac.uk [27] Ainsworth, S.E., Grimshaw, S.K. & Underwood, D.J. “Teachers as designers: Using REDEEM to create ITSs for the classroom.” Computers and Education, 1999, 33(2/3), pp171-188. Available via: http://www.psychology.nottingham.ac.uk/staff/Shaaron.Ainsworth/teachers_as_designers.pdf [28] Ainsworth, S.E. “REDEEM: ITS authoring environments and human teaching.”, Paper presented at the ITS2000 workshop on Modelling Human Teaching Tactic., 2000 Available via: http://www.cogs.susx.ac.uk/users/bend/its2000/ainsworth.pdf [29] Ainsworth, S.E., Williams, B.C & Wood, D.J. “Using the REDEEM ITS Authoring Environment in Naval Training” IEEE International Conference on Advanced Learning Technologies, 2001. Available via: http://www.psychology.nottingham.ac.uk/staff/Shaaron.Ainsworth/icalt.pdf
International Conference on Engineering Education
July 21–25, 2003, Valencia, Spain. 5