multisearchthink-aloudusabilitytest.docVIP

MultiSearch “Think-Aloud” Usability Test April-June 2004 Leads: Kristin Antelman, Karen Ciccone Facilitation and Report: Kelsey Libner, Yinon Bentor Executive Summary 2 Background 3 Summary of Method 3 Results 4 MultiSearch Tasks 4 Likes and Dislikes 6 Mental Models of MultiSearch 8 Findings and Recommendations 8 Appendix A: Summary of Recommendations 11 Appendix B: Method 12 Participant Profile 12 Test Design 12 Discussion of Method 13 Appendix C: Summary of Technical Failures 14 Appendix D: MultiSearch Screenshots 15 Executive Summary A think-aloud usability test of MultiSearch was conducted in April 2004 to determine how effective the product is as a library service. Six undergraduate and four graduate students participated in the test. Participants like the variety of sources offered by MultiSearch, but this test has shown that problems such as technical failures, confusing aspects of the interface, a subject scheme that didn’t make sense to everyone, and inappropriate database selection put considerable barriers between users and the results they are seeking. Furthermore, several users held inaccurate mental models of MultiSearch’s function, often confusing it with the library catalog, but it is not clear that this misunderstanding affected their ability to use the product. Most users were able to select an appropriate group of databases, navigate the interface, and interpret the results successfully. However, searches returned relevant results only about half the time. Eight of the ten participants encountered at least one technical failure in MultiSearch, ranging from minor error messages to very long stalls and dead-end error pages. Recommendations include correcting all identified technical problems, simplifying the additional databases display, providing descriptions of database content coverage within MultiSearch, displaying progress indicators for lengthy operations, ranking results by relevance, moving fielded searches to t