Oct 18 2006
Republican Wave In MD Senate
I predicted last year Republican Michael Steele would win MD’s open Senate seat. The fact is the black community in MD is not going to vote in another lame white guy when they have Steele on the ballot, with the best campaign commercials ever produced. And now we see a poll showing a tie and this race – which means a pick up for Reps (with NJ in most likely):
In an election for U.S. Senator in Maryland today, 10/18/06, Republican Michael Steele and Democrat Ben Cardin tie, according to a SurveyUSA poll conducted exclusively for WMAR-TV Baltimore and WUSA-TV Washington DC. 20 days to the 11/7/06 election, it’s Steele 46%, Cardin 46%.
If there was a Democrat tide Steele would be losing, Lamont would be surging, Allen would be be behind, Menendez would be cruising… Why? Because these are statewide offices where party alliances can get massively diluted or swamped. There is no tide. There are a number of polls with messed up voter turnout models and a lot of conservatives who are not answering the phones for pollsters (and I suspect a lot of supposedly clever liberals pretending they are Reps to pollster so as to bolster the fantasy). I may be whistling past the graveyard, but the I see more energy on the right – but a quiet energy. A determined energy. A subtle energy. While liberals go manic in their anger, everyone else is just going to take care of business.
I have been talking about an iron determination among my friends keeping them bucked up and converting fence sitters. Using words of quiet determination and reminding people of the dirty tricks from pervious elections and how it i s the same as before is slowly working. No matter what happens stay on the issue of security.
[i]Two days before the 2002 mid-term elections, Zogby vs the result.
Mondale was cruising by 5 in Minnesota
Minnesota
Mondale 50
Colman 45
Too bad he lost by 3 points.
Mondale 47
Colman 50
Strickland was a lock by 9 in Colorado
Colorado
Strickland 53
Allard 44
Just two days and a 16 point swing later he lost by 5.
Strickland 46
Allard 51
Max Cleland, what can I say? Result in bold. Are you noticing a pattern here?
Georgia
Cleland 49 46
Chamblis 49 53
My, my .. the beat goes on.
Texas
Kirk 48 43.3
Coryn 49 55.3
Many are familiar with Thune, who didn’t lose by the 5 Zogby projected. How many people stayed home on the margin? Your guess.
South Dakota
Johnson 52 49.6
Thune 47 49.5
Zogby had Lautenberg right in NJ, but under-polled Forester by 6 points.
New Jersey
Lautenburg 54 54
Forester 38 44
Carnahan by 1 but 2 more popped up for Talent and he won. The typos are via Kos.
Missouri
Carnhan 49 49
Tallent 48 50
And going back to Oct. 22, these pearls of wisdom from DailyKos via Charlie Cook:
Cook notes that Dems have a good shot at taking GOP seats in Arkansas, New Hampshire and Colorado, while the race in North Carolina has become unexpectedly competitive…Consider, perhaps the surest takeover bet is Arkansas, a GOP seat. And if the GOP loses New Hampshire, which I believe is likely, then forget it. There’s no way the GOP can take the Senate.
Too bad Dems lost all of the above by 5 – 10 points.
MSNBC says Big Dem Wins Likely. Now where have I seen that before?
[/i]
http://www.riehlworldview.com/carnivorous_conservative/2006/10/a_unique_perspe.html
More proof that the blue tide is crimson–today’s NYT is warning of long lines and Diebold problems and fraud. HEH
Clarice,
You mean the Dems are not quite sure they are going to win in a walk and are setting up their excuses? HEH!
And what about the “exit polls” in 2004? Lynn Cheney later told of ‘keeping the bad news’ away from her husband, the Vice-President, as long as she could. Then, when he saw what the data said about some area (Carolinas, I think) where he just knew better, he said, “No, that can’t be right,” and started checking. She said he knew without a doubt there was something funny about the polls.
Dick Morris also insisted after the election that it had to be deliberate, organized phony numbers turned in from some of the exit poll centers, that it wasn’t plausible for that many different locations to have the same errors.
Early evening in election 2000, I turned off the TV in disgust when Fox (Brit Hume) called the Presidential election (wrongly) for Gore. It was a couple of hours later that I got a phone call from a friend who is by no stretch a political junkie, but she exitedly asked what I though about how the Bush-Gore thing was going. At my disgusted grumbling, she said “No, they took it back, there’s no winner yet…”
This poll thing, I think, is just one of the leftist ploys to suppress GOP turnout. Ignore ’em, the rat b___ards! New motto: Believe nothing a dem says without a certified note from Benedict XVI!
Ignore ‘em, the rat b___ards!
Now there’s a term I’ve haven’t heard in a long time. Excellent advice.
There is a plausible explanation to all this. Let me see if I can describe it without all the math in it. Polls are small samples which people attempt to use to measure a large population. Most scientific polls attempt to keep the variables, the factors that must be used to expand the sample to represent the population, to a small number so that there is little chance the variables push the result off the mark. When we are dealing with a sample size of 500 trying to represent a population size of say 500,000 then 1 percent mistake at 500 can explode (and it is obviously not 100 times or we would see 100% errors).
Modern political polling has gone the other route. They have increased the number of variables (gender, age, education, race, income, etc) which they use to generate a result. This has worked in decades when their was small shifts in public moods. The reason being is the pollsters use historical trends to define all these variables and retain their accuracy.
In 2004 (and earlier) these variables (\”the models\” to those of us who deal with statistics and predictions) began to shift in undetectable ways. And it was a combination of factors such as migrating to the conservative position and pollsters only being able to access smaller and smaller numbers of people (the hang up effect).
The variables gave pollsters the false hope they could adjust, but they couldn\\\’t. That is why they did not see the 1994 wave coming. And some of this is simple denial (which is why I worry I am just not facing the coming Democrat wave!).
But the fact remains, the models (the set of variables) used are crucial to the accuracy. The margin of error use in a poll is only based on sample size, not accuracy of the models. So even the error they claim can be way off. The margin of error assumes the variables are accurate to a certain degree. If they are not, then we see the results we are seeing here, with wild changes in races from poll to poll.
The reality is still this: if Reps sit out Dems could win, if Reps (and cons) vote the polls will mean nothing. Turn out expectations is another variable. All one has to do is defy it.