NC State University Libraries Search Usability Testing

1. NC State University Libraries Search Usability Testing January 2010 ALA Midwinter Conference Josh Boyer & Susan Teague Rector

2. Who We Are • Susan Teague Rector • Web Design Project Librarian • Twitter: seteague • Josh Boyer • Associate Head, Distance Learningand Research and Information Services

3. Today’s Agenda • Background • Usability Test Methodology • Testing Results • Implementation of Search • Search Usage • Next Steps

4. Background • Website Redesign January – August 2010 whose main goal of site design was to emphasize and streamline search across the site • Team of 9 librarians appointed to look at the website’s search environment • Building materials: • Endeca-based catalog interface • Summon • Journal Titles list • site search • “QuickSearch”

5. www.lib.ncsu.edu/search/?q=xml

6. www.lib.ncsu.edu/search/?q=gis

7. How about Tabs?

8. Research Questions • How would users interact with a tabbed search interface? • Would users choose to scope their searches by preselecting a tab before beginning a search or would they conduct all searches in the default “all” tab? • How successful would users be in selecting the correct tab for the task?

9. Search Prototype

10. 2 Search Prototypes Model 1 Tabbed Search Results Model 2 Non-Tabbed Search Results

11. Our Methodology • “Guerrilla” usability testing in the field • Participants recruited in the lobby of DH Hill Library • Testing took place over a 3 hour period • 28 undergraduates, graduates and library staff, participated in a round of guerrilla usability testing for 2 proposed search models • Each participant was asked to complete 2 tasks using one of the search models; many participants volunteered to answer more than 2 questions

12. Who Were the Participants? • 22 Undergraduates • 3 Graduates • 2 Library Staff Members • 1 Non-Traditional Student • Science/Engineering: • Aerospace Engineering • Animal Science/Biology • Biology • Biomedical Engineering • Chemistry • Civil Engineering • Computer Science • Electrical Engineering • Food Science • Human Biology/Nutritional Science • Integrated Management & Systems • Meteorology • Textiles Engineering Represented: • Humanities: • English/Creative Writing • History • Psychology • Business: • Business • Business Administration • 14 participants performed 46 tasks using Model 1 • 14 participants performed 38 tasks using Model 2

13. Observations • In both search models, in well over half the tasks, participants pre-selected a search tab before beginning their search.

14. Observations • Across both search models, in 45 of 60 tasks in which users selected a tab, they selected the appropriate tab for that task.

15. Observations • In both models, users rarely switched silos in the search results once in a silo. Users tended to stay in the same silo and refine searches within those silos.

16. What Worked

17. Problem Areas

18. Other Issues • Even if users executed a successful search, they often did not scroll down far enough in the search results to find the answer • Almost all participants only scanned the first page of search results for a quick answer to the task • Spelling was problematic

19. www.lib.ncsu.edu

20. Search Usage

21. Use of “All” tab (QuickSearch) 8.3% 36.4% 43.8% 2.4%

22. Articles – Full Text Downloads 17.3 % increase

23. Next Steps • Conduct more usability testing with students. • Interview faculty about their research and how they view our new site.

24. Questions? • Thank you! • Contact us: • susan_teague_rector@ncsu.edu • josh_boyer@ncsu.edu