.
If you have heard people (Rs in this case) talking about the pollsters skewing the sample in favor of Obama, here is at least part of what they mean. (For a general description of polls and error see this piece in the NYT.) Some of the polls will show you the behind the scenes stuff if you click on it. This one came from the Wash Post poll in OH.
The main result of the poll reported on RCP was that among likely voters (LV) Obama was ahead by 4.
If you look at the breakdown of Ds, Rs, and Is you get the following:
The debate is over whether that is a reasonable split for Ds and Rs.
At first thought it might seem like it is not. But you would have to know the breakdown of Ds and Rs in Ohio to say that with confidence.
But, the pollsters and Ds say, the real question is how many of each will actually come out to vote in this election. No one knows the future. The pollsters look to the turnout in the 2008 election as being the best guide for what will happen in 2012.
But, say the Rs, there are two things wrong with this.
1. The electorate is not nearly as excited about Obama this time as they were in 2008 and
2. You should use the most recent election which was 2010 which certainly did not have 37-30 Ds to Rs turnout advantage for the Ds.
If you have heard people (Rs in this case) talking about the pollsters skewing the sample in favor of Obama, here is at least part of what they mean. (For a general description of polls and error see this piece in the NYT.) Some of the polls will show you the behind the scenes stuff if you click on it. This one came from the Wash Post poll in OH.
The main result of the poll reported on RCP was that among likely voters (LV) Obama was ahead by 4.
If you look at the breakdown of Ds, Rs, and Is you get the following:
Generally
speaking, do you usually think of yourself as:
Democrat Republican
Independent Other No opinion
9/23/12 OH 35 26 35 3 2
9/23/12 OH RV 35
27 34 3 1
9/23/12 OH LV
37 30 30 2 1 The debate is over whether that is a reasonable split for Ds and Rs.
At first thought it might seem like it is not. But you would have to know the breakdown of Ds and Rs in Ohio to say that with confidence.
But, the pollsters and Ds say, the real question is how many of each will actually come out to vote in this election. No one knows the future. The pollsters look to the turnout in the 2008 election as being the best guide for what will happen in 2012.
But, say the Rs, there are two things wrong with this.
1. The electorate is not nearly as excited about Obama this time as they were in 2008 and
2. You should use the most recent election which was 2010 which certainly did not have 37-30 Ds to Rs turnout advantage for the Ds.
The thing about any kind of predicting is that, unless you are an actual prophet, you must make assumptions. Even if a predictor is 100% politically neutral, the assumptions made can be argued by reasonable people, if you could find any.
ReplyDeleteI often find myself being asked to prepare projected financial statements. And the farther into the future you go the less reliable the assumptions are likely to be. By the same token if you base your assumptions on past results, the farther back you go the less reliable those assumptions are likely to be. So I would tend to agree with those who are critical about using assumptions based on 2008 results when we have persuasive evidence that what happened in 2008 is very unlikely to be repeated.
A competing notion is that 08 was also a presidential election and 10 was not.
DeleteYes, obviously that is true, and should be considered as a factor. But I don't believe that it is a factor that should outweigh all other factors. The history of Presidential elections would teach us, I believe, that it is unlikely that any Presidential election will exactly mirror the last one. The current issues and attitudes are not the same as they were 4 years ago and people know Obama a lot better than they did 4 years ago. They say that familiarity breeds contempt and I can't argue with them (whoever THEY are). On the other hand Obama was called by Jon Stewart earlier this week the "The luckiest sob in the world" for drawing Romney as his opponent. Then he wondered out loud how in the world Romney got the Republican nomination and there followed a memory sequence remembering the other Republican candidates which lead Stewart to proclaim Romney "the 2nd luckiest sob in the world".
DeleteI laugh with Jon, but I don't think he makes any effort to conceal his bias.
DeleteRemember a lot of folks thought that Jimmy Carter was very lucky in getting Reagan as an opponent.
As noted below the MSM is already into the mode of: "lets declare the election over" and get those conservatives to give up.
Polling shows a marked shift in Obama’s favor following 9/11/2012.
ReplyDeleteAssuming (there we go again) that all of the pollsters did not significantly change their assumptions and methodology abruptly on 9/11/2012 then I think the “polling composition argument” used by some Rs is extremely weak at best.
I do NOT think that they changed their techniques on 9-11. So I agree that the bias would not explain the change. However, the argument is that since a bias existed in the samples (before and after 9-11) that the polls have overestimated the Obama vote from the beginning and it is now very close.
DeleteSome would say that it is part of a concentrated effort to declare Romney Dead on Arrival. Notice, in the previous post, how Professor Krugman, a Nobel Prize winner and strong Obama supporter, was willing to believe a major Romney-negative story without checking.