Wednesday, October 24, 2012

The Latest Polls: 'We don't have a clue'

The polls are all over the place. If you average them out it's a statistical tie with Romney holding the thinnest of leads but Obama ahead in enough states that he would still win in the electoral college.


Visibilium said...

John, don't count them and take an average. Examine each methodology. Pew is the gold standard.

M. Jordan Lichens said...

I'm going to make a screen shot of the next sentence as I don't think I'll ever repeat it again. I agree with Vis one-hundred percent. All the same, with the margin of error it does look like a tie at the moment. This is turning into one interesting election

ochlophobist said...

Am I correct in saying that there hasn't been new Pew numbers since Oct 8? I can't find any since then.

Visibilium said...

Och, that's correct. In the meantime, Rasmussen is a good alternative.

Mr. Lichens, the margin of error could be erroneous. The less erroneous the results, the wider the margin of error. I'm laughing as I write that.

While we're on the subject of statistical crocks, I'm tickled by Panagopoulos' flip-flop on Pew/Rasmussen's 2008 efficacy. In his earlier report, Pew/Rassmussen were top-ranked, which corresponded to the relative closeness of their predictions:

A year later, he produced a follow-up study with a greater number of statistical measures. Even though Pew/Rassmussen called the election closest, they ranked somewhat less than tied for first place:

Pew/Rassmussen's efficacy must have consisted in a lucky guess since the predictions didn't conform to the measurements, right?

ochlophobist said...

One problem with all of the polls is that there is no sound method which accounts for the percentage of votes which will be added, subtracted, or otherwise altered by the GOP crony owned electronic voting machine companies. Add that to the equation, and it would seem Obama doesn't have a chance in hell, unless powers that be have decided a conservative revolution will be more feasible in 2016.