| The
intent of this project is to use search term selection decision
as a way to illuminate cognitive structure and explore differences
in the networks between experts and novices. In addition, we were
interested in what other factors might influence search term selection,
possibly overriding the effects of naivete. Feedback effects were
also manipulated. We were able to identify nine distinct search
strategies from the search term data. Results indicated that experts
used what would likely be considered even more efficient strategies
than novices. In addition, feedback that was consistently positive
did override some of the debilitating effects of being a novice
searcher.
The working assumptions in this
project are that search term selection is determined by existing
cognitive structures that the subsequent learning from retrieved
resources will likewise be influenced by the knowledge/experience
that a searcher has available to them. The next step in this project
will be to take the data to the next level by visualizing what
these semantic networks look like so we can begin describing them
using dimensions of dispersion, reach, breadth, associative strength,
and the like. Knowing these, we can better predict what the searcher
is more likely to select for learning in retrieved resources.
These parameters would improve machine-learning algorithms for
query expansion systems and metadata extraction methods.
|
