What? - Collaborative Web Search: Who, What, Where, When, and Why - page 19

information resource for a purpose other than its intended one). While Bates' tactics focus on the
single-searcher scenario, some of them nonetheless recognize the role of the social in information
seeking. For example, the consult tactic involves asking a colleague for suggestions. The brainstorm
tactic involves generating many ideas, which, while not explicitly collaborative, certainly often is. And
the bibble tactic involves looking for ready-prepared bibliographies, or, in other words, reusing ma-
terial already found by trusted others. Bates' inclusion of tactics such as these reveals the importance
of social interactions in everyday information seeking processes.
Twidale et al. ( 1997 ) observed the way students used computer terminals to find information
in a university library, noting several types of collaborative interactions. Based on these observations,
they recommended aspects of information seeking where social interactions could play a role, with
an eye toward reminding the designers of emerging digital library systems of the importance of
preserving the types of social interactions that were possible in traditional library settings. Twidale
et al. distinguished two collaborative search strategies: process-related collaboration (collaboration
on how to find information) and product-related collaboration (where the collaboration involves
exchanging the sought-after information itself ). They also reported several specific tactics people
used to implement these strategies, which are described in Table 3.5.
Researchers have built on these explorations of traditional collaborative library practices by
exploring the tactics people use to search for digital information collaboratively. Morris, M.R. ( 2008 )
surveyed 204 information workers about when they used Web search tools, including search engines
and hyperlinks, to search collaboratively. She identified several specific search tactics where users
engaged in social activities. These included observing another's Web search over-the-shoulder and
suggesting alternate query terms, instant-messaging with other users to coordinate real-time in-
formation seeking in remote scenarios, and explicitly dividing up responsibilities for a search task
among several people and then sharing the results of one's assigned sub-task with the group. Like
Bates and Twidale et al., Morris also found that brainstorming with others, particularly brainstorm-
ing alternative query keyword choices, was a common social interaction during information seeking
tasks.
Morris' survey also identified common tactics in which users interacted with others to share
the products of an information seeking task, summarized in Table 3.6. These included sending a list
of links or textual summary of findings over email or printing them to share with someone in person,
using the telephone to share search findings verbally, turning one's computer monitor or connecting
a computer to a projector to share screen contents with a co-located user, and creating a webpage,
Wiki, or other electronic document to record one's search findings in a format re-usable by others.
Evans and Chi ( 2008 ) surveyed 150 people using Mechanical Turk, asking these users to
answer questions about their most recent information seeking experience. Based on these findings,
they proposed breaking down the opportunities for social interaction during information seeking
into three categories: before search , during search , and after search (in contrast with the two-category
process / product dichotomy used by Twidale et al. and Morris). Their model is illustrated in Figure 3.1.
Evans and Chi then identified several tactics in which their survey respondents used social resources