TruthNews Commentary, January 9, 2008

Hilary Clinton has won the New Hampshire Democratic presidential primary, confounding pre-election polls which showed Barack Obama a clear winner.

National Public Radio, a.k.a. National Left Wing Radio Funded By Your Tax Dollars (NLWRFBYTD) has offered several reasons why the polls were wrong:

- Democrats are lying racists who told the pollsters they would vote for Obama but once in the voting booth found themselves unable to mark the box for the (gasp!) black man

- New Hampshire women voted for Clinton out of pity after she cried at a pre-election rally

- Women are unable to make up their minds

While all of these may sound like perfectly sound reasons to liberal thinkers, we offer a different possibility: polls are inherently inaccurate because the use a sample of the voters to predict the behavior of the entire electorate.

If you sample a hundred people to predict how a hundred thousand people are going to vote, you may find that all one hundred say they're going to vote for Obama. The pollsters would conclude that Obama was going to win 100 percent of the vote. Then these one hundred people all go to the polls and actually vote for Obama like they said they would. Meanwhile, the other 99,900 people all go to the polls and vote for Clinton. Clinton would win with 99.9 percent of the vote, and the pollsters would wonder why they got it wrong. This scenario has low probability, but it is possible, and so are a lot of other scenarios in which the poll doesn't reflect the reality. That's why in the 2004 presidential election, the election results differed significantly from the exit polls.

The ease with which polls can be manipulated is easily seen when campaigns hire pollsters, each of whom produces a poll that shows his candidate is winning or gaining ground. The pollsters then argue with each other about which poll is right.

The inaccuracy of the New Hampshire polls illustrates once again the truth of the adage that figures don't lie but liars use figures.