Notes on Ethnography and Design

CSE 510 -- Winter 1999

Ethnography

• Purpose: understand human behavior through observation, interview and particpation.
• Why is ethnography relevant to design?
  1. Designers create artifacts for work settings -> need to understand those settings.
  2. Technology shapes practice -> designer's world view should not be imposed inappropriately on the users.
  3. Allows us to gain broader understanding of technology in use (see traditional methods)
  4. Joint exploration of technology and work allows both users and designers to participate in new designs.

Guiding Principles of Ethnography

1. Natural Setting
   "To learn about a world you don't understand you must encounter it first hand"
2. Holism
   "Behaviors can only be understood in the everyday context in which they occur."
3. Descriptive
   Descriptive vs. prescriptive approaches -- ethnography is descriptive (Programming language example: building a object-oriented UI in Smalltalk by modifying an existing one)
4. Member's point-of-view
   Get an "insider's" look (see illustrations in paper for worker's vs researcher's point-of-view regarding a typical office work setting)

Observation 1

• Why Observe? Why not just ask?
  ideal vs. manifest behavior
  "What people say and do are not the same thing."
  People have a great deal of tacit knowledge that they cannot verbalize.
• Observational role:

  Unobtrusive Observer (aka Observer Participant)

  Sometimes the ethnographer can be given a culturally appropriate role that allows the ethnographer to hang around. In other settings this can be difficult or inappropriate.

  Participant Observer

  +access, +first-hand experience, +-point-of-view issues, -logistical problems of recording, -problems of "going native"

• In reality, the ethnographer doesn't assume either of the above extremes, but moves back and forth along the continuum between pure observer and pure participant.

Observation 2

• What do we focus on?
  1. Event focus - meetings, seminars, ceremonies.
  2. Person focus - "a day/week/year in the life of..."
  3. Place focus - receptionist's desk, printer room, etc.
  4. Object focus - life history of a document, transaction, etc.
• How do you know you are finished?
  When you're no longer suprised by what you're seeing!

Recording Media/Techniques

• Notes:
  • + inexpensive, interpretive, flexible (sketches, etc), holistic
  • - interpretive, can be hard to reconstruct, "low bandwidth," lots of work for observer
• Audiotape:
  • + inexpensive, "medium bandwidth," less work for observer
  • - limited slice of activity (audible/verbal only), requires some audible actions to be useful, difficult to reconstruct
• Videotape:
  • + "high bandwidth," can make detailed reconstructions (content log), advantages of audio, less work for observer, unobstrusive, somewhat holistic
  • - limited slice of activity, expensive, building content logs can be exhausting, coordinating multiple videos difficult, activity space/lighting might not be suitable for filming
• Event Logs: (computer capture of input events)
  • + "high bandwidth," can make detailed reconstructions, can get good statistics about failures/difficulties, time sensitive
  • - computer only, emphasizes human-machine dyad view -- human is viewed as mere source of input events (dubious politics, IMO), anti-holistic, reductionist

Interviews

• Who to interview?
  Beware of the "well informed informant" problem.
• Guidelines: Mostly these are common sense, but are very important.
  1. Allow respondents to shape interaction (especially early in the investigation).
  2. Don't sacrifice rapport to obtain answers. Remember that some things may be off limits.
  3. Interviewer needs to maintain the attitude that they know less than the interviewee. Remember: you want to learn about the respondent's world (not your view of the respondents world).
• Interviewing will not disclose objective facts -- it is just another source of subjective information.
• Context of interview is important.

Traditional Approaches

• Customer Surveys -- usually v. early or v. late in the process
  Nice statistics about "user community" but there are all sorts of problems with relevance and target. Often, because of the multiple choice nature of the surveys, the answers are constrained in advance.
• Operability Assessments -- usually v. late in process
  "human-machine dyad," lab settings, get performance statistics on (often) very constrained tasks
• Focus Groups Cross section of users/customers.
• Field Trips

Contrasts Between Traditional and Ethnographic Approaches

• Context
  All trad. approaches (except field trips) take place outside the users' work place.
• Focus
  Traditional -> Technology: Is the technology acceptable
  Ethnographic -> Relationship between technology and work/activity.
• User role
  In trad. approaches, there is little room for user collaboration: user is seen as object of study rather than a partner, participant, etc.

Issues

• How do we link ethnography and design? (A few scenarios)
  1. Ethnographer reports findings to designers
  2. Ethnographic study is undertaken by team of ethnographers and designers
  3. Full participation in desing by ethnographers, designers, and users

• Whose "side" is the ethnographer on?

  "In research, an anthropologist's paramount responsibility is to those he studies. When there is a conflict of interest, these individuals must come first." -- American Anthropological Association Principles of Professional Responsibility

  Can an ethnographer supported by a technology company be "both user advocated and purveyors of technology"?

• Who "owns" the result (knowledge) of ethnographic research?

• How much does the presence and actions of the ethnographer observer change the situation? Should the ethnographer try to minimize his or her influence on the situation?

• Ethnography as academic research vs. ethnography in support of engineering design.

  Related technique: contextual inquiry (can be viewed as ethnography adapted for design and engineering)