5 Dec 2002 15:43 AR AR178-PS54-19.tex AR178-PS54-19.sgm LaTeX2e(2002/01/18) P1: IBD
496 OLSON
¥
OLSON
merely to select the appropriate card and post it for others in the cockpit to refer to
as they set their parameters and monitor the descent. The close coordination of the
individual players’ cognitive activities and how they arrange their environment to
coordinate actions with others is the key to claiming that the unit of analysis should
be the whole system, not the individual. Thus, to understand the behavior of the
individuals, one must examine the entire environment in which the activity takes
place. “Such factors as the endurance of a representation, the sensory modality
via which it is accessed, its vulnerability to disruption, and the competition for
modality specific resources may all influence the cognitive properties of such a
system” (Hutchins 1995, p. 286).
This line of thinking has inspired a number of design ideas, notably in the design
of the work environment. Deep analyses of how people use their desks with its
piles of papers and sensitivity to location have led to interfaces that display results
of searches in spatial arrangements reflecting some aspect of the material, either
its date or content. (Czerwinski et al. 1999). Other analyses of how people work
with their whiteboards in both their offices and in meetings led to the development
of electronic whiteboards and clever ways to combine electronic and physical
artifacts (Moran et al. 1997, Mynatt et al. 1999).
Scope of the Theories
Theories and problems in HCI today parallel some of the theoretical transitions
that we witnessed in the 1970s in cognitive science and in the 1960s in verbal
learning. The question is, how much of people’s behavior can be explained by
factors generalizable to all users, regardless of domain expertise or meaning. In
the early 1970s in the area of verbal learning, we broke from studying learning
of nonsense syllables to learning in richer environments in which meaning was
key (Prytulak 1971). In the 1970s there was a lot of work in the spirit of the
GeneralProblemSolver(Newell& Simon 1972), attempting to find out thefeatures
of people’s intellectual processing that were common to everyone, where they
relied only on simple general strategies such as hill-climbing and generate-and-
test. Soon thereafter,countervailing research focused on the nature of expertise and
the specific strategies that experts used rather than generic properties of cognitive
processing (e.g., Ericcson & Charness 1994).
In HCI there is a similar transition from understanding generic behavior(Gestalt
explanations of the understandability of visual layouts, tradeoffs in using mouse-
menu input devices versus learned keystroke combinations) to more knowledge-
centered behavior. For example, a recent study of expert/novice searchers of large
knowledge sources (such as the World Wide Web) showed that both strategies
and successes were highly dependent on domain expertise. The person familiar
with medical information sources was much better and faster at confirming the
appropriateness of a treatment for a diagnosis than the layman and worse at finding
the best price for a consumer product. The search strategies were highly dissimilar
and fully dependent on the domain of expertise (Bhavnani et al. 2001).
Annu. Rev. Psychol. 2003.54:491-516. Downloaded from www.annualreviews.org
Access provided by Beijing Normal University on 03/06/18. For personal use only.