Andrew Fiore
Next steps in online dating: Moving toward social and affective context
Biography
Andrew T. Fiore is studying mediated relationships as a
Ph.D. student with Professor Marti Hearst at UC-Berkeley’s School of Information Management and Systems. For his master’s degree at the MIT Media Lab, he conducted an analysis of the messaging behavior of 57,000 users of an online
dating system over an eight-month period. He also has an undergraduate degree
from Cornell University, where he studied computer science and human-computer
interaction. He has worked as an intern at Microsoft Research on analyzing and
visualizing patterns of interaction in computer-mediated communication. His
master’s thesis is available from
http://web.media.mit.edu/~atf/.
Position Paper
Nearly a century ago, the Sears Roebuck catalog offered almost
every good a person could want, shipped to your doorstep. Online dating
provides such a catalog of potential mates. From the user’s perspective, the
prospects seem bountiful; this surplus surpasses the scale of a singles bar by
several orders of magnitude. However, current systems fail to provide many of
the cues that people use to evaluate potential partners in an offline context.
Although the catalog is immense, it’s hard to tell which items you would
actually like to have. The process of identifying and communicating with a
potential partner is not purely informational; it depends on social influence
and affective processes. Facts about a potential partner matter less than our perception
of the person they describe; this affective, holistic perception depends
not just on facts but on our imagination, our emotional state, and the social
context. The process isn’t fully rational or reducible. Yet online dating
relies on factual search, an oversimplification mitigated only by the sheer
number of possible partners they provide. Volume alone, however, doesn’t help
users find a mate. New technologies and reconfigurations of existing approaches
can give users better tools for this process. Some are possible today; others
will take more time. Here are three major areas for improvement.
Use characteristics that matter. We know a great deal
about what qualities influence attraction and marital satisfaction. We should
revise profiles and matching algorithms to replace superfluous detail with more
germane information. And matching should involve more than just similarity: for
some characteristics, similarity is appealing, whereas for others, opposites
attract. My master’s work examined which characteristics were desirable to
57,000 users on one system; their preferences corresponded closely to those
predicted by psychological studies of attraction offline. This quantitative
work needs as a complement a qualitative exploration of what users value.
Show, don’t tell. Research has shown that people have
imperfect self-knowledge yet do not realize its imperfection. To ask users to
describe themselves and whom they seek, then, is to invite misinformation. We
should present users with ways to demonstrate who they are and what they
like in a partner. Similarly, why not give communicating dyads an activity that
allows them to learn about each other in context instead of describing
themselves over email? (Jeana Frost at MIT is pursuing this agenda.)
Incorporate the biology of chemistry. Capturing the
ephemeral elements of attraction that lurk just beneath our conscious
perception of others remains the most difficult challenge for online dating. We
can judge our unspoken attraction to someone almost instantly face-to-face, but
weeks of online conversation are insufficient to the same task. Pheromones play
some role in what we perceive as “chemistry,” perhaps promoting genetic
compatibility and the potential for healthy children. Whatever the underlying
mechanisms, the experience of chemistry is elusive in mediated channels. Finding
a way to transfer that experience online, or predict what pairs of people will
feel chemistry when they meet in person, would help online dating make the
transition from shallow catalog to powerful tool.
Back to Social Computing Symposium 2005
|
