Cabaniss Usability Study
- IRB approval for the Cabaniss usability study was obtained in May 2010.
Usability Study Flier and FAQs:
- Advertising began on August 6 and ran through August 30, 2010. Ads were placed on the University Libraries main web page, as well as on the student information portal called MyBama and on the Digital Collections web page. In addition, information was emailed to the History, American Studies, and Political Science Departments. These departments forwarded the information to their graduate students and faculty members. Fliers were also displayed in the reading room, and those work work at the reference desk were told about the study and were asked to encourage patrons to take a flier.
- Response to the advertising was fairly positive. We received more volunteers than we needed, but all volunteers were contacted and invited to come and participate in the study, in hopes that enough volunteers would participate to reach our goal of 20 participants. We recruited nearly 30 volunteers before taking down the ads, but due to no-shows, recruited a second time. We finally tested 21 participants, though one set of results was disqualified from analysis due to unwillingness to complete the test.
Usability Study Informed Consent:
- We used the Morae software to perform the usability studies. More information about this software can be found at http://www.techsmith.com/morae.asp.
- Usability testing was completed in September 2010. The study consisted of asking participants to perform a total of 8 tasks, 4 within the Cabaniss collection (http://u0003_0000252), and 4 within a similar collection which was laid out in our "traditional" fashion (The Robert Jemison Jr. Papers Collection: http://acumen.lib.ua.edu/c/u0003_0000753 ).
Usability Study Questions:
After the participants performed these 8 tasks, they were asked to fill out a survey which gave them the opportunity to give us feedback on their experience and their opinions about the interfaces in general and the Cabaniss collection in particular.
Short Discussion on Results
The usability study and its results are detailed in the an article (currently submitted to JCDL 2011) entitled "Low Cost Digitization of Manuscript Collections: Trading Usability for Access?"
One of the more interesting findings was that participants without previous digital collection experience found the finding aid interface significantly easier than those who claimed familiarity with the traditional digital library interface.
Educational experience clearly improved ability for participants to locate content via the finding aid, as did a background in history or familiarity with special collections.
Not surprisingly, non-native speakers had difficulty with both interfaces.
There was very little difference in the apparent learnability of the two interfaces, implying that the finding aid interface may be as learnable as the usual digital library interface (based on item-level description searching). We determined, however, that a true test of learnability should cover a longer period of time and far more queries.
The question arose as to where the boundary is between testing the interfaces, and testing the delivered information?
The majority of (primarily undergraduate) participants preferred the item-level description interface. However, 40% of those with a background in history, and a third of those with a background in special collections, preferred the finding aid interface.
(Previous research has shown that the finding aid as interface to digitized manuscript collections is preferred by scholars and experienced researchers, whereas undergraduates prefer item-level description. Given that our study focused on undergraduates, our results are not surprising.)
We also decided that it is truly not fair to compare the finding aid interface to digital content against the item-level interface for a short-term test. The finding aid presents far more information and context, requiring more time to read and become acquainted with the organization of the content. Once acquainted with a collection, however, we expect that locating appropriate content would be as speedy (or more so) than via searching item level metadata which may or may not contain the necessary search terms.
Research has indicated that presentation of digitized content via the finding aid is indeed the preferred delivery method for scholars. Since this is also by far the cheaper method for digitization and web delivery, we believe this is by far the preferred path for future digitization of large manuscript collections.
Data Tables from Usability Study: (Note: Kevin W. Walker performed the statistical analyses of the results.)