Much of the punditry has an issue with statistics. It’s not that they hate statistics. They would rather pack a room full of people who are very uninformed, very likely, avid debate watching voters and personally ask these clueless voters about their unmade opinion regarding candidates who both have been campaigning for the last year and have been on the national scene for about 5 years. Then they can turn to the real Americans watching and interpret, for us, what real Americans are thinking. I think this is much more fun for them than: MATH!
Silver’s no stranger to doubt and criticism. He even doubts his own model sometimes. But he dismisses this criticism.
“We can debate how much of a favorite Obama is; Romney, clearly, could still win. But this is not wizardry or rocket science,” Silver told POLITICO. “All you have to do is take an average, and count to 270. It’s a pretty simple set of facts. I’m sorry that Joe is math-challenged.”
MSNBC’s Joe Scarborough may not be mathematically challenged, but he’s got a problem with processing and accepting mathematics in narrative busting reports. He had the same issue with October 2012 BLS Unemployment Report that put the number at 7.8% He demanded an explanation had guests on shortly thereafter who explained the report to him (the drop was due to two consecutive months of positive adjustments and real job market growth) and then still said someone had to explain the BLS Report to him. “The labor department needs to explain this to us” or something. The report explains “this” to us. It’s why people write reports. Really it shows that a lot of punditry eats up anecdotes from guys like former GE head Jack Welch who tells bed time stories about the Uncertainty Goblin who eats up all the job openings while Job Creators don golden armor to fight Tax Gargoyles so that we may work for them more. There is one problem: the polling analysis sites are very transparent.
You can go and download data from these guys. There is no magic or b.s. with poll sites that use real regression analysis. They get props by being right, not by being biased whereas Dick Morris, Harold Ford and many other pundits make money being announcers at the horse track with Romney in the pole position after a strong performance at the 1st debate Stakes and Barack Obama “coming from behind” to win the Presidential Derby.
— Nate Silver (@fivethirtyeight) October 27, 2012
The gist of Nate Silver’s modelling is explained here. If you don’t want to click through (you should): It boils down to Nate Silver has built a data model that compares current political data with historical data. That is to say: Silver gathers relevant data from national polls, state by state polls and significant indicators factored in (incumbent vs. challenger, economic climate figures etc). Then he runs a simulation (e.g. 100 people flip a quarter 100 times to find probability of heads or tails) and then the results of the simulation determine the probability of one candidate winning and another losing.70+% chance of getting enough electoral college votes out of that to get to 270.
Ezra Klein defends Nate Silver against criticism:
If Mitt Romney wins on election day, it doesn’t mean Silver’s model was wrong. After all, the model has been fluctuating between giving Romney a 25 percent and 40 percent chance of winning the election. That’s a pretty good chance! If you told me I had a 35 percent chance of winning a million dollars tomorrow, I’d be excited. And if I won the money, I wouldn’t turn around and tell you your information was wrong. I’d still have no evidence I’d ever had anything more than a 35 percent chance.
I’ve used many of these sites before to talk down fellow liberals even though he’s not the only guy to go to. I came to them by way of fantasy football. By “guys to go to”, I mean guys who run sites that have solid records in last four or five elections of being as close as you can be with their projections. It’s not because they are liberal and we agree (they aren’t) or because we are basement dwelling bloggers or geeks who can type with two hands and drink craft beers or have been to San Francisco or live in a walk-able city (they aren’t). It’s because when I use their data, it’s soundly supported and explained with math and collected, verifiable data.
Here are the sites that aren’t fivethirtyeight.com run by left libertarian Nate Silver and have Obama beating Romney and all have developed prediction models and historical results vs. their previous models:
- electoral-vote.com – a left leaning American abroad, a great site I’ve been using this since 2006
- election.princeton.edu – another solid site, just newly on my radar as of this election
- electionprojection.com – a right leaning, born again christian runs this one.
- Intrade let’s people put their money where their mouth is and Obama has a 66.5% chance as of this posting.