Oct 20 2006
Dems Notice The Poll Bias
As I argued when I looked at the RCP top 40 House races there has been a stubborn 3-5% bias against Republicans in the polls. This tends to give the Democrats a real false sense of success – as we saw in 2004. The problem is compounded by the fact pollsters (at least on exit polls) tended to get their responses in city centers, missing what was occurring in less populated areas. Ohio was a classic example of this mistaken measurement in 2004. The bias is due to the pollsters, not anything else. They have tried to refine their models over the years by adding detailed subgroups of voters, but they do not refine their measurements to match the models (how many men, women, reps, HS graduates, etc). The bias has been there and can be removed – but it takes money. To make sure your sample is representative of the historic dynamics for all the groups you are trying represent requires time and money.
Now the blabbering talking heads on the left are finally noticing the issue, with some hillariously naive attempts to understand statistics and math. I swear someone should do a poll on math skills between Reps and Dems, because Dems always fall for obviously bogus information because they just don’t test it with some basic math. Like the idea of 700 Iraqis dying per day due to the insurgency. Anyway, Molly Ivins is quite depressed today about the news:
The argument now is that Ds have a seven-point structural deficit going into any election. I see the problem, I just have no idea what the actual numbers are.
…
“As journalist Hendrik Hertzberg (of the New Yorker) notes, if you treat each senator as representing half that state’s population, then the Senate’s 55 Republicans currently represent 131 million people, while the 44 Democrats represent 161 million people.”
OK, we all know about the small-state advantage in the Senate. How did the People’s House get so far out of fair? Paul Krugman explains: “The key point is that African-Americans, who overwhelmingly vote Democratic, are highly concentrated in a few districts.
“This means that in close elections many Democratic votes are, as political analysts say, wasted — they simply add to huge majorities in a small number of districts, while the more widely spread Republican vote allows the GOP to win by narrower margins in a larger number of districts.”
This is the NY and LA effect. If you poll a lot there the world looks mighty blue with Democrats. But smart pollsters know better than this. That is why this is a specious argument meant for those who need a simple explanation to grasp. The truth of the matter is in 2004, even with these centers of liberalism, Bush took had a larger number of voters. It was a 52-48 context, with less than half of elligible voters participating. And that is the other rub. Turn out. In the general population the Democrats are enjoying the remnants of a nostalgic echo from their past, when they were a party of ideas. But the general population doesn’t vote regularly at all. Ivins marks this down to racism, but it is a trend in all demographic groups. So who comes out in off year elections? The die hard, well informed foot soldiers of democracy (since some of us do not belong to either party). And in that group the first hand knowledge and experience of American politics blunts all the Democrats ‘messages’. To get that population you have to stand for something. And the conservatives are not bowed down by the avalanche of biased reporting coming out of our useless news media (time to just shutter up these palaces of ignorance). But the reality is starting to hit the Dems, as Ivins explains:
I’m the one that keeps beating the Washington press corps about the head over how out of touch it is. I’m the one who has been insisting there’s a Democratic tide out here, and that the people are so far ahead of the politicians and the media it’s painful to watch.
So how come I’m not thrilled? Because I watched this happen two years ago — same rejection of the Iraq war, same disgust with Bush and Co., same understanding Republicans are for the rich, period, same polls showing Ds with the lead going right into Election Day. And the same geographic gerrymander and same wall of money in the last two weeks. I’m not close to calling this election, and I’m sure not into celebrating anything yet.
Same old tired Liberal fantasy world which fewer and fewer Americans subscribe to. The reason liberals are losing is they only tell us what they want reality to be for them – not for us.
First Cup 10.20.06…
……
Maguire has just demonstrated this is false.
Clarice,
Yeah, I saw his post. The bias is real between polls and results, you posted on that yourself here. What is false is the explanation that Alterman came up with. What I am wondering is whether all this Rep talk of doom is real, or establishing expectations so when the Dems do not get what they want, it will destroy them confidence wise.
Molly Ivins is really depressed… Racism, gerrymandering, to much last minute money and lying. Sounds like she is getting herself ready for a defeat!
Why would anyone expect the Dems to understand the polling bias? This is the ’60s generation of which I was youngster.. bystander. They react emotionally first and then when wrong spend their time arguing til you tire and walk away.. This is the generation that “got by” with excuses and never, ever where really tested or challenged. So, why is it a surprise that they have not ever understood that 2004, 52-48 is a crushing defeat in national politics? Thus, they will never be able to identify their mistakes, make amends and get on the road to recovery.
More on poll bias over at Redstate
http://www.redstate.com/stories/elections/2006/a_bit_of_sunshine_for_your_morning