11.10.2006

Thin-Slice Forecasts of Gubernatorial Elections

In his recent best-seller, "Blink," Malcolm Gladwell references a
study in which students were asked to watch a half-hour clip of a
professor and then guess what that professor's student rating would
be. They were something like 95% accurate. The study then takes
another pool of participants and asks them to watch a five-minute clip
and do the same, and their results are something like 92% accurate.
The study goes as far as five-second clips, and still the accuracy is
90% or more. Gladwell calls this "thin-slicing," or the ability to
extrapolate to great accuracy despite have just a sliver of
information.

This month, the National Bureau of Economic Research published a
timely paper called "Thin-Slice Forecasts of Gubernatorial Elections,"
in which participants watch a ten-second clip of two gubernatorial
candidates and then are asked to predict the outcome of the election.
Their accuracy is better than sophisticated economic forecasting
models and on par with what you'd predict if you knew incumbency
status. Even better (or worse), accuracy deteriorated when
participants were given information about the issues and positions
candidates were campaigning on.

Is this what politics has become in a sound-bite world, that I can
tell you who's going to win by looking at a picture and peeking in for
ten seconds? Or does this demonstrate the profound ability of people
to "thin-slice" and, out of a cacophony of information drill down to
what's actually going happen? Is it bad that we so easily jump to
conclusions and let those conclusions orient our opinion of someone or
something? Or does it just mean that you never have a second chance
to make a first impression, so make sure to add some polish to your
polish?

Frankly, I don't know, and my head spins the more I think about it.
But I thought this all was very interesting, and while I don't know
what the implications are, I know that there are implications.

Post a Comment