Lies, Damn Lies, Statistics...and Voter Polls

September 8, 2008 9:40 AM

Margins of Error We can't seem to let the future alone. Even though we often get predictions about it so wrong. Because, as Niels Bohr once said, "Prediction is very difficult. Especially if it is about the future." What are the origins of political polling (beware of awful interface design)? And how is political polling evolving?
posted by jeanmari (5 comments total)

"Marge, it takes two to lie. One to lie and one to listen."
- Homer Simpson

also:

"Phfft! Facts. You can use them to prove anything."
- Homer Simpson
posted by blue_beetle at 10:02 AM on September 8, 2008

The Bohr quote is disputed (but funny nonetheless).
posted by AwkwardPause at 10:26 AM on September 8, 2008

Thanks for this. I can use the David Moore interview in class.
posted by wittgenstein at 10:29 AM on September 8, 2008

I believe the correct quote is "Facts are meaningless. You could use facts to prove anything that’s even remotely true!"
posted by fings at 11:29 AM on September 8, 2008 [1 favorite]

Something that bothers me about the Slate article (and about other similar claims about markets accurately predicting some event) is that markets don't actually "predict" anything. They only reflect the relative likelihood of an event occurring based on the available information at that time. Sometimes, unlikely things just happen without enough warning for the market to catch up, as was the case when Obama unexpectedly lost in New Hampshire.

If the markets were trading, say, totals for a pair of dice, 12 would always have a low price because it's relatively unlikely, but sure enough it would actually happen sometimes (once out of 36 rolls, on average).
posted by piers at 8:37 PM on September 8, 2008