Kathleen Fear: “User understanding of metadata in digital image collections: Or what exactly do you mean by ‘coverage’?”, in The American Archivist, vol. 73, number 1, 2010, available as a PDF, is a study worth reading.
This study looks at the usefulness of Dublin Core metadata in digitized image collections from the user’s perspective. Are there elements that users don’t find useful? Do users feel useful information is missing? Are labels and elements provided for users arranged in a way that makes sense to them? This paper uses a survey, focus groups, and search/usability testing to gain insight into the kinds of information nonexpert users rely on when searching for images and to identify the vocabulary that best expresses that information.
Here’s why I find this paper invaluable: the users involved in the study were undergraduates or other non-expert users, and they are the ones with the most exposure to sites like Amazon and Google and thus rather likely to turn their back on libraries. It is vital to study their perception of library catalogs and metadata since their approach is different from experts who are usually much more accustomed to the way metadata is currently presented in libraries.
The study focuses on Dublin Core, which is said to be rather simple – too simple for some. But for the participants the information contained in the DC metadata was sufficient and there was nothing they needed to know that wasn’t in the metadata.
The fact that the resources at the centre of the study are images coincides with my recent interest in these types of material. Those of you working with digital libraries in general, and with digital image collections in particular, may also appreciate the author’s insights. Creating metadata for images and searching for images are quite different from textual documents, but I only mention this in passing, because more detail is beyond the scope of this post. I believe, though, that several conclusions from the study are valid for “traditional” library catalogs as well.
One finding that struck me, but also confirmed what I had suspected, is that “the interface … can hinder the perceived usefulness of some elements, especially when an abundance of text caused the user to skip over everything”. This is certainly a peculiarity of image searching: the image itself (or its thumbnail) seems to be the primary source of information, and especially when there is a lot of descriptive text, users don’t read it. This correlates with the fact that a lot of library users don’t go beyond the brief display. Interface design however can affect the usefulness of textual metadata. Are there lessons to be learned for bibliographic catalogs?