Information Technology Reference
In-Depth Information
comments were positive, with eight of ten users saying they enjoyed the
experience and would play again.
Pre-Use questions focused on the size and composition of participants online
and offline social networks, and how they use social networking software. All
participants thought they knew their social networks well, and "usually" become
friends on Facebook with people they meet offline, and the vast majority "almost
never" meeting people offline they meet on Facebook. Estimates of
communication by strength and type, including email and Facebook, showed little
correlation among users.
Most users completed five to ten surveys in the allotted ten minute timeframe.
From these results, a desired participant response of 200 questions was selected
under the assumption that this selection would take approximately 15-20 minutes.
After using the software, eight of ten users said they would use the software
again, with the remainder questioning the utility of rating one's friends, a
manageable loss percentage for the study. Approximately half of respondents
requested the ability to compare their results with their friends. Users found the
selected data points interesting, and performed exceedingly well, with many
reporting 100% accuracy, and indicating that the questions needed to be harder.
From these user comments, a series of modifications were made focusing on an
improved participant experience and improving data collection methods. As many
users requested, comparisons to one's friends was added in the form of the
Stats
page, containing groupings of “People Who Know You Best” and “People You
Know Best.” Some users mentioned that the pictures contained in the quiz were
too small to be helpful, and their size was significantly increased. The selection
algorithm was also modified to better incorporate the presence of associated
downloads in the dataset. Some users were also found to log in once, complete a
quiz, and log out assuming that those twelve questions were sufficient. Therefore,
a progress bar was added to indicate to users how much data remains to be
collected from both the individual user and overall.
From this study, the selection algorithm and user interface were vastly improved.
4 Results
Participant responses indicate that users have proportionately more friends of
weaker connection strength, are more accurate at gauging connection intensity
when the connection is strong and they claim to know both participants well, and
connections within cliques are perceived more accurately than those outside.
These results confirm the primary hypotheses being tested:
(1) online social networks contain larger numbers of weak connections that
represent an increased diversity of information flow,
(2) connections rated to have higher intensity will be perceived more accurately,
(3) increased user closeness correlates with improved perceptual accuracy about
social networks, and
(4) connections within a completely connected subcomponent, or clique, will be
more accurately perceived than those between cliques.
