This project, undertaken in collaboration with OCLC, aimed to investigate the potential role of recommendations within WorldCat, the publicly accessible union catalogue of libraries participating in the OCLC global cooperative. The goal of the project was a set of conceptual design guidelines for a WorldCat.org recommender system, based on a comprehensive understanding of the systems users and their needs.
Taking a mixed-methods approach, the investigation consisted of four phases. Phase one consisted of twenty-one focus groups with key user goups held in three locations; the UK, the US, and Australia and New Zealand. Phase 2 consisted of a pop-up survey implemented on WorldCat.org, and gathered 2,918 responses. Phase three represented an analysis of two months of WorldCat.org transaction log data, consisting of over 15,000,000 sessions. Phase four was a lab based user study investigating and comparing the use of WorldCat.org with Amazon.
Findings from each strand were integrated, and the key themes to emerge from the research are discussed. Different methods of classifying the WorldCat.org user population are presented, along with a taxonomy of work- and search-tasks. Key perspectives on the utility of a recommender system are considered, along with a reflection on how the information search behaviour exhibited by users interacting with recommendations while undertaking typical catalogue tasks can be interpreted.
Based on the enriched perspective of the system, and the role of recommendation in the catalogue, a series of conceptual design specifications are presented for the development of a WorldCat.org recommender system.
