Why? - Collaborative Web Search: Who, What, Where, When, and Why - page 68

relevance improvements, systems might also be identifying situations in which it would be valuable
to suggest more explicit collaborations among that user group.
Of course, privacy concerns are important when automatically bringing together or suggesting
collaborators. It may be that for personal or sensitive topics collaboration should not be suggested,
or care should be taken to ensure the collaboration is anonymous or with people far outside the
searcher's social sphere. Morris et al. ( 2010c ) asked participants to describe why they would choose
not to answer status-message questions. Many people indicated private topics were a demotivator,
although some indicated that they might respond privately to such inquiries. This suggests systems
intended to support collaborative search over personal topics should be clear about how information
is shared and with whom. Some topics in particular that may be considered too private for suggesting
collaborations with previously-unknown users or with users having certain types of relationships such
as professional ties, include religion, politics, sex and dating, personal details about friends or family,
money, and health ( Morris et al. , 2010c ).

6.4 WHY: CONCLUSIONS AND FUTURE DIRECTIONS

Future collaborative search tools can benefit greatly from understanding why people collaborate. We
explored several reasons why people engage with others beyond situations in which they are already
explicitly cooperating on a shared task. People also turn to social resources when such sources
broadly share their interests in a topic or have special knowledge of a topic, as well as when a topic
is challenging to address through traditional Web search. Social motivators, including a desire to
connect with others, trust, fun, and reward systems also incent group information-seeking activities.
Identifying when users are investigating a topic well-suited to collaboration and considering
how implicit social search techniques like collaborative filtering, search communities, or groupization
might be useful for identifying groups of people who could benefit from more explicit collaborations
are intriguing areas for future research. Such systems might help users form new collaborations by
identifying trustworthy friends, experts, or even unknown people with similar interests. Understand-
ing how such systems might preserve privacy, as well as the role that social rewards and reputation
systems might play to incentivize engagement, is a rich area for further exploration.