As a distance learning institution our provision of library resources and services is almost entirely online, so a good discovery experience is essential. We wanted to know how our students really use discovery tools. Do they treat library discovery tools like Google? Do they use all the options to refine their search? Do they want all the sophisticated features that librarians and academics care about?
The Open University comprises a large and varied student body of more than 200,000. These students are scattered throughout the UK and over 13,000 are overseas. We used online meeting software to enable students to participate in our research from the comfort of their own homes or workplaces, using a familiar computer. This has allowed students to take part from all over the UK and even from as far afield as Australia. This approach was tested in 2013 as part of an exercise to understand student requirements for personalisation of library services.
Starting in November 2013 we carried out a series of interviews, observation exercises and testing to find out how students use existing discovery tools, what they think of them and what they want from library search.
The research has been carried out in three phases. In all three phases students were given a selection of search scenarios to use which were taken from activities in OU modules and from our search logs. The scenarios included finding a known item from a reference, a known resource such as a database or dictionary, or undertaking a search for materials around a given topic as though preparing to write an assignment. A literature search helped us to understand common difficulties with discovery and identify examples of different approaches to providing user-friendly interfaces.
- Phase 1 established student requirements for library discovery through observing students undertaking search tasks in different search tools and asking for feedback. 8 different search tools were tested in this way, including Google Scholar and two tools that the OU library had developed. 54 scenario/tool combinations were tested.
- Phase 2 used wireframes of a series of search and results pages to test different options based on the student requirements identified in phase 1 without taking the time (and cost) of building a working prototype. This meant that we could rule out features that students didn’t find helpful at an early stage and concentrate on building the features that are most likely to be well used.
- Phase 3 tested a prototype search tool which was developed as a proof-of-concept to test with students. This enabled us to develop a checklist of features which can be used to undertake an options appraisal to decide whether they can best be met with an out-of-the-box discovery implementation, a custom search interface (possibly based on the search prototype) or another alternative.
Unsurprisingly we observed that students' search strategies and expectations are often based on their experience of using Google and they expect discovery tool relevance ranking to work in a similar way. They want the search interface to be clean and uncluttered as they find many of the features in discovery tools confusing and distracting. The majority of students only turned to advanced search after they’d tried a basic search and found their results were unsatisfactory. Post-search filtering was more common that pre-search filtering, but only used in fewer than half of the searches they were asked to do. Only use a small selection of the facets available for refining results are used. Instead they generally alter their search terms, even including limiters such as content type or publications dates, expecting the discovery tool to work like Google. Autocomplete suggestions for the search box were used by around half of the interview participants and considered familiar by all of them. Students were strongly in favour of having a simple way of saving results they wanted to use or come back to, such as a virtual bookshelf.
Keren Mills is Digital Services Development Officer at the Open University. The full findings of this research will be presented in session B203 at the Internet Librarian International conference in October 2014.