Franklin Fact Archive
January, 2000
February, 2000
March, 2000
April, 2000
May, 2000
June, 2000
July, 2000
August, 2000
September, 2000
October, 2000
November, 2000
December, 2000
January, 2001
Back to Franklin Facts homepage.
Back to TV12
|
MONDAY DECEMBER 4 - ELECTIONS AND METEOROLOGY
As I was watching the Election Night television coverage back on November 7, particularly
how the networks had trouble predicting who would carry Florida, I kept thinking: "I know
how they feel."
Here's why. There's a data-collecting organization called Voter News Services that conducts
exit polls on election day, and then supplies the networks with that information. The
networks examine the data and project winners. In this past election, Voter News Services
polled about 70,000 voters nationwide, less than one-tenth of one percent of the 100
million or so people who voted. When you think about it, that's not a very big sample on which
to base forecasts.
Here's the weather connection. Good observations, and lots of them, are essential for a
good weather forecast. But the truth is, the atmosphere is huge, and we only have
observations of important quantities such as temperature, pressure, and humidity from a
very, very small sample of that air. From those relatively few observations, we
construct our "best guess" of the actual atmospheric conditions. It's sort of like
calling elections based on just a small sample of voters. And it's that estimation
over vast reaches of the atmosphere that often leads to misrepresented storms and
consequently inaccurate forecasts.
|