Replies

I find the somewhat triumphalist discussion of poll numbers recently on FR amusing.

Only a few weeks ago, when Romney was way behind, the most common reaction around here was that the polls were worthless.

So now that we like the results they’re suddenly accurate?

I realize, of course, that they may still be understating the degree of support for Romney for the same reasons they were criticized before.

BTW, I wish polls weren’t as accurate as they are. Polls have taken a great deal of the meaning and interest out of elections.

Sherman,

Good Morning! Actually none of the poll junkies at FR have changed our tunes.

We all believed that the polls UNDER reflected Gov Romney’s strength by 4-5% points as most polls were using 2008 turnout samples

So, we were always compensating for that in our analysis. What is happening now is that DESPITE the pro Democrat weighted polls, Gov Romney is either showing as tied or ahead.

For example, take today morning’s WSJ/NBC poll. It shows Gov Romney tied at 47% with President Obama. In my opinion, based on their poll size and poll history that poll undercounts Gov Romney’s potential vote share

In the end the only poll that counts is the one in the ballot box. On Nov 6th we will find out who was right. Were pollsters like Rasmussen right (who only show a 2 point gap between Gov Romney and President Obama) or was Gallup right in capturing the enthusiasm and turnout of Republicans

Then again, by tomorrow, President Obama and Governor Romney could be tied in Gallup :)

This is certainly a strange polling season and we are all trying our best to understand the math, stats and analytics that are driving these polls.

Only a few weeks ago, when Romney was way behind, the most common reaction around here was that the polls were worthless.

First of all, Romney was never "way behind" at any point in this race. As for the polls being "suddenly accurate" now that Romney is ahead, well that's a function of the election being only about a couple of week away.

The polls are always more accurate as we close in on Election Day. This is because pollsters need to stay in business for future election cycles so any manipulation needs to go away.

We are still not quite at that point yet. There is still some special sauce "cooking" these polls. For example, using a D+8 sample for Ohio? Ridiculous. But over the next 15 days, you will see that Ohio sample be more like D+3 and Romney will be 5-6 points ahead on their final Ohio polls. Count on it.

Look at any modern presidential election and you will find that nearly all major pollsters (Gallup, Rasmussen, etc.) are surprisingly accurate with their final polls. However, go back a few weeks and they are all over the place. Remember in 1980 that September polls showed Carter beating Reagan handily - does anybody believe that? Or in 1996, when Clinton was ahead of Dole by 22 points in September.

So yes, the polls of September were garbage and as we head into late October, the polls are suddenly showing a Romney "surge". In my opinion, there is no Romney surge. The pollsters are simply beginning to come into line with reality so their reputations will be secure for future elections.

Only a few weeks ago, when Romney was way behind, the most common reaction around here was that the polls were worthless.
So now that we like the results they’re suddenly accurate?
I realize, of course, that they may still be understating the degree of support for Romney for the same reasons they were criticized before.

Actually none of the poll junkies at FR have changed our tunes. We all believed that the polls UNDER reflected Gov Romney’s strength by 4-5% points as most polls were using 2008 turnout samples.
So, we were always compensating for that in our analysis. What is happening now is that DESPITE the pro Democrat weighted polls, Gov Romney is either showing as tied or ahead.
Agreed.

BTW, I wish polls weren’t as accurate as they are. Polls have taken a great deal of the meaning and interest out of elections.

I think we need to pull the plug on all the uncertainty to the maximum possible extent. And the way to do that is to institute a system by which facts and logic are placed on prominent display and allowed to throw political sophistry into bold relief.
If you look up “political sophistry,” what you should find is a video of Joe Biden trying to laugh the facts and logic of our situation out of the debate - with cooperation by the moderator. The opposite of that situation would be

multiple debates, perhaps one a week from the end of the conventions to the last week of the campaign. Thereby limiting the impact of a “gaffe” and maximizing the possibility that the public will understand the truth.

Said debates to be televised as a condition of retaining FCC licensure - removing the journalists’ negotiating strength to control the format.

Debaters should not be limited to being “talking head” programming, but should be allowed and encouraged to use a web site to document all statistics, charts, and graphs they want. They should be able to put a crawl under the screen advertising their web sites. Starting in October, all the facts appealed to should be sourced on the web well before being used in debate.

The moderator function should be replaced by a simple chess timer by which a debater turns off his microphone and his "time allotment remaining” countdown clock, and turns on those of his opponent. Each candidate should control his own camera (and perhaps, his own half of the viewers’ TV screens). There should be no restriction on the ability of a candidate to access advice from his seconds via a laptop or a tablet.
The purpose is not to obtain a “gotcha” but to make the facts and logic being appealed to explicit. This runs directly counter to the interests of journalism - and of the sophistry-dependent Democratic Party.
I want the process of electing our officers to be boring - because a reasonable selection, rather than the risk of a President Obama, is foreordained.