Nov 3, 2012

The Corporate Media Narrative if Romney Loses

Well, three more days, and it'll be done. National polls have been tight, with a tiny lead for Romney, but the state polls, especially the battleground states have been breaking largely in Obama's favor. Multiple poll aggregation sites, such as Nate Silver's 538, Sam Wang's Princeton Election Consortium, PuffHo's Pollster (all declared Obama supporters) and Real Clear Politics (Republican leaning site) have this.

To summarize as of 11/3/2012
[UPDATED: Obama's win had little to do with Sandy]
  • Nate Silver has Obama winning 305 electoral votes, 50.6% of the popular vote, 83.7% chance of winning.
  • Sam Wang has Obama winning 323 electoral votes, with a 98.0%+ chance of winning.
  • Pollster has Obama winning at least 277 electoral votes
  • RCP has Obama winning 290 electoral votes (no tossup model), with 47.4% of the popular votes, (Romney has 47.2%)

I'll leave the cranks aside, because they have Romney winning. This narrative does not apply to them.
But if you listen to the news, or watch TV, there are pundits who are screaming that it is a toss-up, and that the race is close as hell.  Despite all the polling evidence to the contrary.  Well, this is to be expected because pundits get paid for rhetoric and drawing viewers, not for objectivity. Joe Scarborough went the Full Neanderthal calling Nate Silver ignorant because he had Obama's chances rated a week or so ago at 73.6 percent.

50.1 percent was the number Joe pulled out of his posterior to counter the aggregated poll based forecast.  How did he come up with it?  Well, he is a pundit, and here's his pundit-ization:
Nate Silver says this is a 73.6 percent chance that the president’s going to win. Nobody in that campaign thinks they have a 73.6 percent — they think they have a 50.1 percent chance of winning.
…. Anybody that thinks that this race is anything but a tossup right now is such an ideologue [that] they should be kept away from typewriters, computers, laptops, and microphones for the next ten days, because they’re jokes.
Got that?  Nate Silver is a joke, because Joe talked to a campaign.

Dick Morris, the ever-wrong pundit had declared it to be a landslide on the boob tube, but on his site, he was at a more modest 301 electoral votes.  On November 1.  A day later, this was Dick updating his projection.

Now imagine what happens to these pundits should the results fall close to the guys doing the math, doing scientifically grounded statistical analysis?

They'd look like fools. That would matter to Joe, but not to Dick, he has not been correct in a while, and it does not matter on FOX "News" anyway.

Now note that none of these statisticians do any actual polling. Which means that if they are wrong, they have bad models, which they need to fix. Sam Wang was spot on in 2004, and off by just one electoral vote in 2008, Nate Silver was the same in 2008. What will be of concern is the actual polling, and all the pollsters who actually polled the states would be wrong.  That would be scary because polling is a well established scientific process, time-tested and seldom that wrong.  Consider this: yesterday there were some twenty states polls in which Obama was ahead in seventeen or so, and tied in all but one of the others.  If Romney ends up winning those states, what does it say about the polling?

But what if they are right, as it looks likely that they'll be?

Then the punditry profession takes a huge hit. All the armchair expertise, pulse of the campaigns, the gut feelings--all are shot to hell in a Eastwood wicker chair or something.

There are vested interests here. So watch for this to be played in the media in the days to come, should Romney lose.

"Were the experts wrong? Tune in to Hannity tomorrow to find out!"
"Does Joe Scarborough owe Nate Silver an apology? Not according to this expert who breaks it down"

and so on.

And how will they wiggle out of this?  By regurgitating various versions of the following narrative:
  1. After the third debate, the race was a toss-up, literally would have been like Florida 2000, except in more states.
  2. Then Sandy hit
  3. Obama pulled away because of Sandy.
  4. Nate Silver/Sam Wang cannot prove that Sandy did not clinch it for Obama.
  5. Also Chris Christie
Note that the above is as devoid of any evidence as the claims of the 50.1 percent number that Joe extracted out of you-know-where.

When this happens, and it is already beginning to take root, see here, here, here, and here for starters, remember these two things:
  1. Specificity is not certainty
  2. Estimating a probability is not making a prediction
I'll expound on those themes in my continuing series on the abuse of statistics, but if you want a different perspective right now please look at these excellent articles [updated: 11/4/2012] by Mark Coddington, David Roher, and Zeynep Tufekci.

Meanwhile, my prediction which was based on "gut feeling" like Joe's was that if the unemployment rate moved down from 8.2 percent to 8.0 percent or below, Obama would win.  If not, Romney would.  Then I had this other "gut feeling" which said that if Romney got the SuperPAC funding he was targeted to (around 1.2 billion) then he would win.

Well, I was surprised that the unemployment rate actually fell lower than 8 percent, but more so that Obama too raised a lot of money. Accordingly I did a mea culpa and am throwing my lot with the math dudes. Sam Wang's history is too persuasive to ignore.

Who did I vote for?  Gary Johnson.

In conclusion, I have a request of my readers: if you find this theme proliferating after the elections, please cite a link to this post in your responses if you choose to challenge the perpetrators.

1 comment:

  1. People are always so worried about the Citizen's United decision ruining the US election process, but the fact of the matter is, the news is what is ruining the US election process. As you say, they're not in business to be objective, they're in business to make money on a story. The story "Batshit lunatics aren't around anymore because no one gives them any credibility, the end" is pretty low on the thrill-o-meter scale.