Draft  and Revise Test

• We began planning in the Fall of 1998 and we determined the purpose and desired outcomes of testing.
• We spent a lot of time focusing on the best ways to get useful and productive feedback.
• For example, our initial drafts were testing people's research skills, instead of getting useful information about the site.

Test the Test

• Do a trial run on a couple of users (we used student employees).
• Evaluate the effectiveness of your test and look for redundancies and irrelevant questions.
• See the kind of data you will get and to determine if this kind of data will be useful for your needs.
• We learned we had to mix up our questions and shorten the length.
• Edit. Edit .Edit. Edit. Edit.

Revise Test and Plan for Data Analysis

• Think about your particular user community and how best to assess their needs.
• We decided to split our usability study into two different tests in order to get the widest array of feedback.  
• I did the HTML and the basic javascript to launch the pop-up window
• Systems staff wrote a CGI program in that captured the data to a delimited file for processing, which allowed us to collect the data and have a great deal of flexibility for exporting the data into various software packages.
• Data Collection Questions:
  • Kind of Data?
  • Amount of Data?
  • How are you going to analyze the data?  

Conduct Test

• We settled on spring timeline that would let us conduct the test at a peak time for library research -- the two weeks after Spring Break.