Full Thesis: here
Separate files:
- Acknowledgment
- Chapter 1 - Learning Objects & Learning Object Metadata
- Chapter 2 - Structural Interoperability
- Chapter 3 - Actual Use of Metadata
- Chapter 4 - Usability Aspects of Learning Object Metadata Tools
- Chapter 5 - Tracking Contextualized Attention Metadata
- Chapter 6 - Conclusions and Future Research
- References
Preface:
Reusability of learning objects is expected to noticeably elevate the creation of online learning material, and several repositories have been developed worldwide to collect large sets of reusable learning objects. In traditional learning environments lessons are designed for a specific group of learners in a particular context, at considerable cost and effort. In the reusable learning object approach, a small-sized learning object may be created by one person, indexed into a repository and then found and reused, individually or by combining it with other objects, in another context by any number of people. This can significantly reduce the time and effort needed for the creation of learning content.
However, creating instructional material from learning objects on a large scale requires access to large collections of learning objects. Metadata are used to describe learning objects in order to enable finding relevant content, using metadata schemas (Like the IEEE Learning Object Metadata standard – LOM). Metadata schemas also enable the exchange of learning object metadata between repositories. Repositories that conform to the same common metadata schema can exchange metadata instances between them. In this way, it becomes possible to create a course on, for instance, ‘Global Warming’ by combining objects (e.g., images, narrative texts, slides and audio and video clips) from more than one interconnected repository.
There are several problems that obstruct effective and efficient reuse of learning objects in education and training [Duval and Hodgins, 2003]; among the main problems (see section 1.6):
- Enabling large scale reuse requires the availability of a vast amount of learning objects.
- Introducing and finding relevant learning objects in learning object repositories is not straightforward. This is a barrier for collecting large sets of learning objects in each repository.
The above two related problems beg for an answer to the following three questions:
- How can access to large sets of learning objects be enabled?
- Why is introducing and finding learning objects not straightforward?
If manual indexing of learning object metadata is a bottleneck, what other metadata can be collected and how? This thesis contributes to this research area by empirically studying the way end users interact with learning objects and their metadata. In this way, we contribute to find out whether learning object metadata tools suit user needs. In addition, studies like ours provide rich information that contributes to the improvement of the learning object tools, based on the analysed user behaviour and feedback.
This thesis is organized as follows:
In chapter 1, a review of some relevant concepts of learning objects, learning object metadata, relevant technical standards and related research challenges are presented. This chapter also introduces the ARIADNE system and the IEEE Learning Object Metadata (LOM), which are the context of my research.
Before we study how people really use metadata, it is necessary to understand how metadata elements of different learning communities, commonly organized in what is known as application profiles, are organized and how they can be mapped into a common schema.
Chapter 2 provides a general background on challenges involved in mapping metadata between diverse metadata schemas. The steps involved in building application profiles, including the selection of metadata elements, selection of relevant value domains and definition of the importance (optional and mandatory, etc.) of profile elements, are also explained. Concerning the mapping of application profile metadata into common schemas, the possible types of mappings are introduced. Situations where the semantics of metadata information might be lost or preserved are discussed throughout the chapter.
Chapter 3 investigates the actual use of metadata elements and values by both indexers of and searcher for learning objects. In this chapter, metadata instances are analysed. In order to analyse the behaviour of users who search for learning objects, the log files of searcher activities are analysed. In this study, the elements and values that are most often used by indexers and searchers are identified. Statistical correlations between the uses of metadata elements are determined. This helps to understand the way users perceive the concept of metadata when introducing or searching for learning objects. A comparison between searcher and indexer activities is presented. Findings of this study are compared to findings of related work from other domains like internet browsing and digital libraries.
Chapter 4 studies the usability of metadata indexation and search tools. In this study, usability sessions with representative participants are conducted. The emphasis is on user performance and the quality of provided metadata information. Increasing user motivation and satisfaction is also discussed. Results reveal that the current learning object indexation and search tools are more adapted to the metadata standards than to their end users. More interestingly, human indexing of learning objects does not guarantee a good quality of metadata.
Chapter 5 focuses on how we can capture data about the usage of the learning object in different contexts. The main concern is capturing and analysing the interactions of users with the object, trying to find out why they do what they do and when they do it. Related initiatives are concerned with the collection of usage data (called also attention metadata) for one system. This chapter introduces our Contextualized Attention Metadata (CAM) framework and schema for tracking user attention in different tools. CAM streams are collected from server-side sources like digital repositories and client-side sources like office software and web browsers. These attention data may be used to enable users to find relevant information and objects based on their previous activities or activities of similar users; using social recommendation techniques and user profiles. In the CAM approach, detailed metadata about how the object was used, who used it and in what context, are collected. This information can contribute to generating learning object metadata automatically. A case study that shows the flexibility of the approach is given.
Chapter 6 highlights the main conclusions for this thesis and points to directions for future work concerning the empirical investigation of the learning objects and learning object metadata.
Understanding the behaviour of a user is a continuing challenge, because every person is unique in every moment of his life. This thesis explores the use and usefulness of learning object metadata and associated tools for end users. In addition, this thesis presents the novel idea that user behaviour, or attention, needs to be collected and related across system boundaries. This enables drawing a clearer image about the user goals in order to help find relevant information.
