Currier, Sarah; Barton, Jane; O’Beirne, Rónán; Ryan, Ben (2004)
Quality assurance for digital learning object repositories: Issues for the metadata creation process
ALT-J, Research in Learning Technology, Vol. 12, No. 1, pp. 5–20
Related Topics: Learning Management Systems
Review by: Reichert, Raimond (2005-01-03)
Metadata has become an important issue with regards to learning objects and repositories for learning objects. There has been much work on metadata standards such as IEEE LOM and much discussion on how to assure metadata quality. However, the process of creating the actual metadata has largely been ignored, and is often trivialized as being “straightforward” for the content authors where it can not be generated by the computer. The authors challenge this assumption and use three case studies to explore the various problems in creating metadata; in the following, we briefly review one of these case studies.
The first case study illustrates the problem with metadata creation. A taxonomy was developed to classify learning objects, since no appropriate taxonomy was readily available. To investigate the use of this taxonomy, the project team agreed upon ‘ideal’ classifications for eight learning objects, assigning a total of 35 classifications to the eight objects. Six users of the project’s users community were introduced to the taxonomy and then asked to classify those eight objects. The result was quite disillusioning. For example, only about 50% of the ‘ideal’ classifications were agreed upon by more than half of the users. The was a very significant variation in classification which lead to many inconsistencies which in turn drastically decrease the value of the metadata.
The authors suggest that a collaborative model for creating metadata may be most appropriate: The content authors provide metadata such as their name and title of resource; a metadata specialist then checks these metadata for accuracy and adds the classification metadata. This approach improved results in two of the case studies presented in the article.
The article addresses the important issue of metadata creation, challenging widespread assumptions about how straightforward that process is. It also points out that with metadata, the devil is indeed in the detail, a problem often conveniently overlooked. The questions raised by this article should be more widely investigated by the learning objects and content management community.