What? - Collaborative Web Search: Who, What, Where, When, and Why - page 18

well (17% of questions). Although the respondents in Morris et al.'s survey were not actively engaged
in collaborative searches, their findings suggest that the question topics and types users attempt to
address through social means, such as asking one's social network or posting to a Q&A site, might
be ones that are in need of more explicit collaborative search support.
The topics of collaborative search tasks differ somewhat from the topics people typically use
search engines for. A 2004 study of America Online's query logs ( Beitzel et al. , 2004 ) found that
the most popular query topics were shopping (13%), entertainment (13%), pornography (10%),
computing (9%), health (5%), travel (5%), games (5%), and home (5%). Many popular search engine
topics, such as pornography and health, are highly personal. A study of the questions people pose to
their social networks ( Morris et al. , 2010c ) suggests that those topics, along with religion, politics,
dating, and financial issues, were ones on which people may be particularly careful about whom they
collaborate with.
Via query log analysis, Broder, A. ( 2002 ) identified three kinds of common intents people have
while searching: informational (intended to find a piece of information), transactional (intended to
perform some action), and navigational (intended to find a particular resource). Collaborative search
intents can be viewed within this taxonomy as well. For example, while students tend to collaborate
mostly on informational searches, seniors and new immigrants also collaborated with other, more
technically skilled searchers, on transactional (e.g., paying bills) and navigational tasks (e.g., finding
sites with job postings) ( Amershi and Morris , 2008 ).

3.2 COLLABORATIVE STRATEGIES AND TACTICS

Although Web search is typically thought of in terms of entering a keyword into a search engine,
Web search behavior is usually much more complex, encompassing a larger process of formulating
and refining an information need, selecting appropriate resources, and evaluating and using the
information one has found. Social interactions can play important roles at each of many stages of an
information seeking task.
When searching, people employ a number of different search strategies . A search strategy is
plan for how the searcher will go about finding the information that is being sought. For example,
Martha's initial strategy to learn more about asthma after her diagnosis was to browse quickly
through many diverse online asthma-related resources to get a big-picture understanding of the
disease. Search strategies can vary greatly, ranging from orienteering to the information target using
clues from the information environment, to teleporting directly to the target by fully describing it
up front ( Teevan et al. , 2004 ).
Bates, M. ( 1979a , b ) introduced the notion of search tactics , or specific actions that assist a
user in implementing a search strategy. Entering a query into a search engine, for example, is usu-
ally just one tactic in a larger search strategy ( Teevan et al. , 2004 ). Bates, M. ( 1979a , b ) defined 29
search tactics, including ones that involve thinking about how to implement a search strategy (e.g.,
weigh : make a cost-benefit assessment of potential actions), formulating an information need (e.g.,
reduce : subtract terms from a query), and using the information resources found (e.g., stretch : use an