Subject: UWired Usability Committee Meeting Minutes - 5/22/00


              UWired Usability Committee

Next Meeting: Monday, June 26, 9am-10am
              C&C Technical Communications Group, 
              Room 015L, Mary Gates Hall
              
-----------------------------------------------------------------------------
              
May 22, 2000 Meeting Minutes

Attending: Stacy Waters, Tom Lewis, Scott Macklin, Carol Robinson, 
           Geri Faris, Rick Ells, Tracy Wagner
Location:  Center for Advanced Research Technology in the Arts 
           and Humanities (CARTAH)
           Room 35 Thomson Hall
           
Guest:     Geri Faris
           Human Subjects Review Coordinator
           phone: 616-2345
           email: gfaris@u.washington.edu
           web:   http://depts.washington.edu/hsd/


Is Usability Testing Subject to Human Subjects Review?
------------------------------------------------------

Maybe - depends on what you are doing and how you are doing it.

Human Subjects Division of Grants and Contract Services reviews 
research involving use of human subjects (either directly or 
through records or other data). In addition to health and safety 
aspects, the UW as an institution has an interest
in protecting the privacy of human subjects.

Usability testing may seem outside of the scope of human subject 
review:
  - Usability testers view themselves as testing the product, 
    not the person
  - Usability testing aims at quality improvement or quality control
  - Often circular: develop, deploy, evaluate, develop, deploy, etc. 

When does something quit being development and start being research.
 - Its research when data becomes generalizable, it goes beyond the 
   department
 - Its research when there are research funds involved
 - Its research when people are getting research credit

You should communicate with Human Subjects when your work becomes 
research or when it involves collecting data where participants 
can be uniquely identified. Human Subjects review can be at any 
of three levels:
  1. Exemption - exempt from institutional review,
     - No risk, no one's privacy will be invaded
  2. Minimal risk
     - Some risk, requires documentation and discussion of methods
     - Use of video from which subjects can be identified 
       moves the project over to minimal risk, by definition
  3. Full committee review
     - Significant risk to health, safety, or privacy of subjects
Minimal risk and full committee require full application

Suggestions on usability tests
  - Want link between data and person's identity to be destroyed 
    as soon as possible - may seem innocuous, but need to look 
    at it institutionally.
    = Go through video and pick out relevant chunks and delete the 
      rest
    = Get a release from the subject to keep the chunks
    = Have a good process for getting consent - the subject needs to 
      know the purpose of the study, procedures, how long, what their
      risks are, how long their name will be associated with the data, 
      identity protection procedures, whether recognizable images 
      or other data will be kept
    = Accepting payment does not necessarily imply consent. 
      Spell it out.
  - Have solid and cogent justification for keeping any data 
    or tapes that links to individuals. Do you really need such data?
  - Choose camera angles that do not show faces 
  - when in doubt, call Geri Faris at Human Subjects
      
Compensation
  - Whatever method of compensation you use, it should not be 
    coercive
  - Pay needs to be commensurate with expectations to compensate 
    for time and trouble.
    = often pay students less than would pay professionals.
  - Need to be compensated all the way along as work progresses. 
    You shouldn't withhold payment until the entire research 
    project is completed.
  - "You may have already won" is considered coercion
  - "Lottery" (if you take a usability test you will be entered 
    in our lottery to win $100) not a good idea

Recruitment problems often result from three things:
  - Coercion
  - Conflict of interest
    = Unequal power relationships have potentialfor coercion
      o Teacher/student relationship
      o Supervisor/staff relationship
      o Subject must have a genuine opportunity to opt out
    = It may not be in the potential subject's interest to participate
      no matter how much you want them to.
  - Confidentiality
    = Has confidentiality been breached?
    = Just because a subject has participated in your study does 
      not mean you can make their name available for other tests. 
      Get their permission before passing on their identity.


 |- Rick Ells - 543-2875 - rells@cac.washington.edu - Rm 301 4545 Bldg -|
 |-                http://staff.washington.edu/rells/                  -|