Data versus intuition in predicting the election outcome
Daniel Patrick Moynihan is reported to have said, “Everyone is entitled to their own opinions, but they are not entitled to their own facts.”
However, I fact checked this quote and discovered there is some controversy over who originally said this as well as what Moynihan exactly said. Sigh.
So, as it turns out, maybe everybody can have their own facts. Which is apparently what Republicans believe about Democrats. And vice versa.
Nevertheless, I persist in believing there are facts, and I regularly read many of the fact checking sites, though, and I admit this contradicts my premise, I think some of these sites try so hard to appear impartial that they miss the truth. WWDD? (What would Diogenes do?)
You have your PolitiFact.com, FactCheck.org, OpenSecrets.org, Snopes.com, TruthorFiction.com and HoaxSlayer.com. OpenSecrets looks mainly at political money and HoaxSlayer at scientific myths.
But it appears that HoaxSlayer is the work of one guy, so it’s not exactly a jury opinion. PolitiFact is run by a Florida newspaper, so (irony warning) how reliable could it be.
Aaaaarrrgggh. Maybe there are no facts.
But polling data. Surely it’s factual. But the polling data about the upcoming presidential election at first blush seems contradictory with some showing the presidential race a dead heat, others shown one candidate or the other in the lead. On the same day.
You have pollster.com, fivethirtyeight.com, and RealClearPolitics.com. All these sites analyze polling data and try to separate the wheat from the chaff. Other non-data analysts say Romney has momentum and will win.
Psychologists have an ongoing battle with one faction that believes clinical judgment trumps a cold empirical analysis and another faction that believes the opposite. This is called Escoffier versus Betty Crocker…the famed French chef who worked without a recipe versus the queen of the stringent recipe.
Regarding the polls, data analysts say the nation is evenly divided on the upcoming presidential election but that the president has the electoral college edge. Statistics guru Nate Silver gives Obama a 75% chance of a win on Tuesday.
As you know, others disagree, citing Escoffier-like intuition about the mood of the nation.
I’m going with Silver over Escoffier.
But the matter will be settled on November 6th.
Unless it’s not.
Mark Hastert
6 months, 3 weeks agoSome of the coverage of polls appears to be for dramatic purposes to make the horse race more exciting.
Silver, Sam Wang, Votamatic all use models to crunch terabytes of polling data and all have cime under criticism mostly from the right for being skewed but the interesting phenomena is that all of them from 538 to Intrade predict the same outcome. Next Tuesday there is as much on the line for statisticians as for the political parties.
Yanwen Xia
Overland Park
6 months, 3 weeks agoSometimes I thought we have to trust the intelligence of the majority of voters, that is, we trust that people still remember how we got into this huge economic mess in the first place and trust people will not want to go back to that time.
If that trust fails, no matter what the outcome is, I truly believe as the nation we have become wiser after the two wars and would not go back to the Bush-Chenney time. Nothing drains the nation’s resources (human life, too) more than these wars.
Plus, we have to believe the end of election is the beginning of another battle, as neither side will stop fighting with the end of election.
JR Beillenhouser
6 months, 3 weeks agoAs usual, a liberal assumes that he knows the facts. I’m an engineer by training, and I never trust anyone to do the research for me, I do it myself. I look at data and then analyze it, it has nothing to do with intuition. You believe Silver, who is a partisan. He is taking garbage data in and producing garbage data out. Silver did great predicting the 2008 election. He faltered in 2010. This election cycle will either show him to be a great statistician or just a hack who happened to be lucky once.
Why will Romney win?
http://www.gallup.com/poll/158399/2012-electorate-looks-like-2008.aspx
(Under demographics of likely voter)
Therefore samples of +5 to +11 for democrats are complete BS. The sample size for a correct poll should be at least even. Many pollers continue to use a +8 Dem from 2008.
So when a poll comes out like this one http://www.nytimes.com/interactive/2012/10/31/us/politics/31poll-results-documents.html?ref=politics where it is +8 Dems and has Obama up 2 points in virginia, it is more likely that he is 6 down. (p 16)
Silver takes every poll and coverts them to +8 as well to simulate 2008. But this is not 2008.
FYI, Rasmussen is +3 dems.
Early polling is going Romney’s way. Obama needs to get a big push in early voting to withstand election day, where republican voters come out. His numbers vs 2008 are way down.
Romney is winning the independent vote by double digits. This group decides the election. That Quinnipiac/New York Times/CBS News poll listed above shows Romney with a 21 point advantage with independents in Virginia.
Where the candidates appearing and where are they putting their money is a key to determining what their internal polling is saying. In 2008 Obama had the swing states in hand and started to increase his influence to states like Indiana. This election cycle, he is being forced to states such as Iowa, Wisconsin, Michigan and Pennsylvania. It was expected that he would have these states well in hand. Romney, on the other hand, is trying to increase his lot in this states that should be blue. In addition, he is polling well in these states. He will not win all of them, but I expect that he will win Iowa and one more.
In the next few days, you are going to finally see what the rest of us already know, as pollsters will now be concerned with their reputations and will try to get as close as possible to a correct result of the election. Obama is in big trouble.
This election will be either a close win by Obama or a blow out by Romney.
Come election day, you may just find out that you didn’t have the facts, but rather the intuition.
George Harris
Kansas City
6 months, 3 weeks agoJ.R. you could well be correct. If you are, the pollsters won’t be able to save their reputations, and you’ll be able to make a ton of money in the next election!
George Harris
Kansas City
6 months, 3 weeks agoJ.R., here’s another reason to be skeptical about poll results: http://truth-out.org/news/item/12418-political-polling-is-no-longer-meaningful
William R. Nelson
6 months, 3 weeks agoI got my onions; my lantern; and I am so ready to vote. - Diogenes
Mark Hastert
6 months, 3 weeks agoI see this AM that Nat Silver has challenged Joe Scarborough to a $1,000 bet, proceeds to the American Red Cross. No word yet from Sam Wang at Princeton, Votamatic, In-Trade odds or the others. It should be noted that none of the above are saying Romney can’t win, only that Obama has better odds. If they are right it will only reinforce the perception that conservative can’t do arithmetic.
Gee Joe, kill the messenger….
Phil Cardarella
6 months, 3 weeks agoThere are such things as facts.
Unfortunately, we have a singular ability as a nation to deny those which are inconvenient — especially those that are inconvenient for the rich.
Berkeley is names after a famous philosopher/bishop, who faced with the philosophical musing of those who deny the existence of the material world, kicked a rock (possibly at a student’s shin) stating, “Behold, I refute it thus!”
Hurricane Sandy is a rock directed at our collective shins.