Cognitive Artifacts in Complex Work - Ambient Intelligence for Scientific Discovery - page 163

PubMed or MEDLINE exclusively. These two observations indicate highly stable
patterns of information use over time.
Information artifacts are not merely analogous to the cognitive work of re-
search, they are the result, target, and product of cognitive work. Scientists are
immersed in information artifacts. They often scan or read over a dozen articles
a week, run numerous searches related to experimental work, and may simulta-
neously work on research papers and grant proposals while running experiments
or managing lab work. While not obvious from the view of the lab bench, a
life sciences lab is a dynamic, highly productive information environment. But
from a production perspective, much of the information work is churn, exten-
sive and repeated reading to draw specific information objects from the context
of research articles. This difference between artifact and information object is
central to information use in discovery.

4.1 Research Method

Information tasks and artifacts provide a “trail” of cognitive work, matching task
and cognitive artifacts (e.g. search results, articles, abstracts) with participants'
representations of goals, intentions, and observed practice. Cognitive distribution
of information tasks was studied using several methods unique to information
behavior research. The unit of analysis was the research project activity, with
data collected and coded by individual tasks and joint information practices
by project. Information tasks were examined over the time period of identified
research projects. Collaborative processes were described in semi-structured in-
terviews and specified in response to questions in task diaries. Multiple accounts
were obtained and compared.
The studies were conducted as cognitive ethnographies, observing informa-
tion behaviors as empirical actions in the work setting and identifying cognitive
artifacts used in information-driven practices. Field data collected included video
and audio recorded observations and interviews, information task walkthroughs,
and diaries recording detailed information tasks. From a total of 32 life scientists
(molecular biologists and pharmacologists), 20 graduate students and faculty sci-
entists kept electronic diaries for 10 days each, recording their information tasks
as they occurred. Across the diary sampling periods of 3 weeks, over 90% of
the recorded information tasks originated from the demands of experimental
research projects.
Multiple research methods were employed to gather field data and analyze
findings. Drawing from cognitive ethnography, observations followed the par-
ticipant's natural course of work practice. The likelihood of obtaining ecologi-
cally valid data was enhanced by triangulating other methods to acquire specific
records of information use (within participants and between sample groups).
Task diaries were used with 5 faculty and 15 graduate students, gathering
detailed information behaviors on a daily basis for up to 10 response days over
2-3 weeks. The purpose of the diary was to sample actual information tasks
as they occurred during everyday practice. By distributing the diary activity
across a range of participants, different information tasks were described for self-