Wednesday, 16 November 2011

'Users do not understand the difference between querying and searching': ReScript querying issue 1




Learnability, Memorability


Users found the blue circle and red triangle symbols and accompanying text confusing. Some tried to click the symbols, expecting them to be interactive. Even after learning what the symbols and text meant with reference to searching, users continued to find them confusing, and reported having to think about what they meant each time they conducted searches.

Impact severity



Replace the current search legend with a breadbox with a link to a Super Search form which will pre-populate the current query.


Users were often unable to understand exactly what the two search options (blue circle and red triangle) offered, making the search process slow and clumsy.
Although both symbols were clearly distinct, users failed to associate them with the relevant action

Quantitative measure

[Using the alumni of Oxford university]: You are looking for records on Exeter College. Where would you click ?

Actual question

You are looking for all records on Magdalen College. Where would you click?

Initial click test result ('before')

September 2011, 89 responses.


Development change

Redesign using the 'calendar' function, a device familiar to historians, containing links which toggle the highlighting of matches in the article. When highlighted, each match also becomes a ready-made hyperlink to the site search engine [e.g.].
The new calendar function sits in the first sidebar - 4 college entries have been selected and relevant matches within the article body are highlighted accordingly

Follow-up click test result ('after')

November 2011, 330 responses.



Although there were clear groupings of clicks in both the before and after screens, the redesigned screen has more which are correct (the calendar option in the first sidebar and the link at the bottom of the page). Groupings do appear over other incorrect instances in the article body (using the red herring 'Magdalen Hall' as opposed to 'College'); perhaps that might be alleviated by varying the highlight colours. As the quantity of responses varied heavily between the tests (almost 4:1 higher the second time around), it is not surprising that there were more random clicks in the second test.

Feedback from interviewees regarding our initial design, which treated XML marked-up text as a ‘database of words’, and used UNION and INTERSECT queries rather than straightforward searching, prompted a radical overhaul of our vision for the service. The new version focuses more on analysing the particular article on screen at a given moment, and utilises functions which we believe will be more intuitive for historians to use. We look forward to further feedback from users in our second tranche of interviews.

Bruce Tate, Project Manager
Donna Baillie, Project Officer

No comments:

Post a Comment