So, the NYT has an op-ed up about how Hillary beats McCain, but McCain beats Obama. And they have scientists to back it.
Sadly, this goes to show that people with english majors should not attempt mathematics. As is so often the case when reporters do science, the author completely misses the point the scientists were making. This is how the editorial descibes the findings.
This conclusion comes not from wishful thinking but from a new method of analysis on the statistics of polls that has been accepted for publication in the journal Mathematical and Computer Modeling. The authors, J. Richard Gott III, a professor at Princeton, and Wes Colley, a researcher at the University of Alabama in Huntsville, are not political scientists. They are astrophysicists. And one of the tasks of scientists is to clarify the apparent complexity of the universe by using the language of mathematics.
After the break, I'll discuss a bit about how the author is not fluent in the language of mathematics.
What the author describes is an interesting study from the 2004 election:
Here’s what they discovered: in swing states, the median result of all the polls conducted in the weeks prior to an election is an especially effective predictor of which candidate will win that election — even in states where the polls consistently fall within the margin of error.
This method provides a far more accurate assessment of public opinion than most people’s politically informed commentary. In the 2004 presidential election between John Kerry and George W. Bush, many political analysts said the race was too close to call. But when Professor Gott and Dr. Colley applied the median method in 2004, they correctly predicted the winner in 49 states, missing only Hawaii.
Really, I can't say that is a shocking result. You collect as much data as you can, and on average it comes out closer to the truth than any model built on a smaller piece of data, or opinion.
But we now follow this up by ignoring one of the main pieces of the study.
That remarkable success left me wondering what result this method would give if I applied it to the 2008 presidential race.
I'm sure you could, that would be the point of the study.
So I examined the past six weeks of polls...
I followed the simple rules established by Professor Gott and Dr. Colley: in states in which a poll has not been taken, you give that state to the party that won it in 2004. You do the same for states where the median poll is a tie.
Well, no, I'm affraid you didn't follow the rules established by the good professors. You see they said; median result of all the polls conducted in the weeks prior to an election.
You see, we are not in the weeks leading up to an election, so their study says nothing predictive at all. Throw into this that we are; on one side, involved in a hotly contested primary, and in the other on a settled nominee. Saying that you are following the rules on this is like saying, "I prepared the solution the same way as the professors, only instead of using sulphuric acid, I used bathing salts, and instead of radioactive isotopes, I substituted cheese curls.
In the weeks prior! How can you miss that? It is the main supposition in the whole study. But no, he takes the last six weeks of primary polling—and some old polls from 2004 and finds panic! Doom! Hillary was right, she wins and Obama fails.
He ends with two questions, which I will attempt to answer here:
Two questions arise in the face of this result. Whom should the Republican candidate prefer to run against to maximize his party’s chances of retaining the White House? And what does it say of the Democratic delegate selection system when its winner would lose the presidency if an election were held today, yet its loser would win it?
Answer 1: Seriously, who cares? I know this whole election thing is confusing, but it is Democrats that get to pick the Democratic candidate. Sure, I would like to win, but I'd like to win with a candidate of my choosing who is going to do the things I want done.
Answer 2: It might say something if the election were held today, but since it is not, never has been, will not in the future ever be held on the day a contested primary ends, IT A PRETTY FREAKING USELESS QUESTION!
In fact, it says nothing predictive whatsoever.
Just because you add a scientists name to dumb logic, it doesn't make it any less dumb. If anything it makes you look worse to those who realize you have not correctly understood what the scientists were saying in the first place.
Come back and show me how this study looks mid-September and we will talk.
[Update]I appolgize to english majors everywhere. I should have noticed this;Neil deGrasse Tyson, an astrophysicist and the author of “Death by Black Hole and Other Cosmic Quandaries,” is the host of “Nova scienceNOW.” Thus showing that just because you understand science, doesn't mean you understand science--apparently. The op-ed is still dumber than my poll.