Oct 14 2008

Garbage In, Garbage Out

Published by at 9:04 pm under 2008 Elections,All General Discussions

The hysterically looney CBS News/NY Times poll shows why math is not for everyone. As a political veteran said recently the public doesn’t have wild gyrations in mood – especially not in a week.  The tell tale bogus factor in this poll is where they put all their big shift, and it is the views of ‘independents':

10/6

Republicans: Obama=7% McCain=87% 
Democrats: Obama=87%, McCain=9% 
Independents: Obama=39%, McCain=49%

10/13

Republicans: Obama=10%, McCain=83%  (-3% for McCain)
Democrats: Obama=88%, McCain=8% (-1% for McCain)
Independents: Obama=51%, McCain=33% (-28% for McCain)

Does anyone with more than 2 IQ points to rub together think one third of the independents shifted their opinion in one week? If so please contact me, I own a bridge in Brooklyn I will sell you dirt cheap.

Addendum: Folks, I need to point out that if 28% of the independents moved, a hefty chunk of the Reps would move as well.  Since there was no movement at all this data is garbage. There is no way a third of the indies move without pulling at least 10% of the Reps with them. 

Update: Check out the fun Allah Pundit is having on this one over at Hot Air. I am sure Geraghty will chime in soon also.

20 responses so far

20 Responses to “Garbage In, Garbage Out”

  1. grumpyguy says:

    I’ve been tuning in everyday to see your analysis of the polls.

    Allah over at Hot Air posted this CBS/NYTs poll and I headed right over here to see if you had seen it!!!

    This was taken from a comment over at HA concerning this poll, what do you think?

    “his new CBS News/New York Times poll, which is being touted with a 14-POINT LEAD for Obama, up from THIREE POINTS in the same poll a week ago . . . IS A FRAUD!

    A peek into the “internals” of the Poll (that is – where the numbers come from) shows this fraud:

    Total Respondents 1070

    UNWEIGHTED WEIGHTED

    Registered Voters 972 902

    Democrats 371 338

    Republicans 295 263

    Independents 306 301

    *Effective Likely Voters 699

    *Every registered voter is included in the likely voter model, and is assigned a probability of voting, which is used to calculate the likely voter results. The sum of these probabilities is the effective number of likely voters.

    – – – – –

    So… according to the data CBS/NYT is using, not only have they heavily oversampled Democrats over Republicans, but they had hideously oversampled Independants.

    PLUS… they have defined ALL registered voters and being LIKELY voters, which any reputable pollster will tell you is nonesense.”

  2. AJStrata says:

    Grumpy,

    It is a pathetic joke for the Obamabots to drool over.

    Clearly this is some of that new fuzzy math they teach in grade school at work

  3. daralharb says:

    CBS/NYT’s latest offering and similar outlier polls may be a joke, but I have lost my sense of humor. Obama’s partisans believe these polls as some might believe scripture, and they expect nothing but a smashing landslide victory. To the extent that the election produces anything less – even if it results in a narrow Obama victory – these true believers will not credit the results. If McCain should win, their denial is likely to fuel a nearly boundless fury, possibly a violent fury. Disappointed as I may have been, on several levels, with the current administration, I am grateful that national security and domestic order will be in its hands until January 20, 2009.

  4. MerlinOS2 says:

    The other thing it does is slide into the mix with all the other polls and when the averages are figured for all the polls the Obama margin will appear much stronger than it really is.

    There are so many bogus polls with crappy internals right now that should be thrown out on pure ethics grounds from the polling industry that are skewing the whole situation so bad that few can rely on them to any extent.

    The only upside is that if the Obama Bots believe they have it in a walk maybe it will be harder for them to get their follower to show up to the polls.

  5. bush_is_best says:

    Who cares about polls. If you say they are inaccurate, then we believe you, they are. If you say the race is closer than all those statistical models say it is, then ok, you’re right… no argument from the left. You are correct to be questioning the numbers, and the fairness, and the random sampling… sure, probably alot of misinformation there. Good thing we look at these polls with about the same believability as we do scripture… (rough guideline with serious deviance from actual reality)

    Its more an indicator that yeah, we have a chance to change the representation of our country from republican back to democrat. We are displeased with the current situation and embarrassed of our leadership. We’d rather not continue on that course. The polls are merely an indicator that some of our efforts are being realized and that we should continue. The Liberals have a solid change of winning, lets hope they keep up the good work and finish strong!

  6. archtop says:

    “CBS/NYT’s latest offering and similar outlier polls may be a joke, but I have lost my sense of humor. Obama’s partisans believe these polls as some might believe scripture, and they expect nothing but a smashing landslide victory. To the extent that the election produces anything less – even if it results in a narrow Obama victory – these true believers will not credit the results. If McCain should win, their denial is likely to fuel a nearly boundless fury, possibly a violent fury.”

    But what can you do? They can say anything they want (free speech). If they want to say Obama is 14 % points ahead, no one can stop them. The question they must ask themselves is if everyone else is saying 3 – 10%, will they look pretty foolish? Of course they will! They obviously don’t care anymore. Moreover, when the polls start to tighten up as they always do before the election (professional pollsters want to be close to the actual result, after all), then Obama will appear to have fallen like a stone – not the momentum you want going into election day!

  7. [...] AJ Strata explains…garbage in, garbage out. Math isn’t for [...]

  8. crosspatch says:

    Rasmussen was unchanged today, as it has been the previous two days. 5 point spread, and that is after Rasmussen increased their poling sample of Democrats.

    The CBS poll is designed to discourage Republicans from voting by giving the impression that it is a lost cause. It is designed to get Obama elected by giving Republicans a psychological path to apathy.

    The last time I got an opinion pollster calling me, several months back, I told her I was going to vote for Obama. She said “oh, good!” which lead me to believe it was some Democrat outfit doing the polling. I said something along the lines of “Yeah, I am a life long Communist and it is good to see someone from such strong socialist stock finally making it to a position of power. Besides, the Socialist Workers Party has endorsed him. Too bad he has to cover as a Democrat”. She sort of went apoplectic and didn’t really know what to say after that.

  9. daralharb says:

    Indeed, Archtop, I have no concern with what Obama’s more blinkered partisans might believe or say: freedom of belief and expression is in fact our law and tradition. My concern at this point is for public order if these partisans, on November 4, should be disappointed in their expectations. That’s why I am glad that the machinery of government will remain in the current administration’s hands until January 20 of next year.

    As to what I can do, it is no more than what I have done and intend to do. I will vote for John McCain on November 4. Between now and that date, I will devote what time I can to working in his campaign. I will speak to as many people as I can about the issues in this election, and above all about the overriding issue of the character, political style, background, beliefs, and associations of Barack Obama.

    More than that, no one can do.

  10. ShoreMark says:

    CBS TV in the NY market has been pushing excited blurbs all night long (what can I say, I like the Tuesday nite programming on their channel) about the 11PM news that will “tell you about dramatic changes in the polls that are shaking up the presidential race.”

    Yawn. They pull this same crap every 4 years. They’ll follow up with excited exit polls showing Obama up by double digits, just like they did for Kerry — we can only hope Obama does as Kerry did in the real world of ordinary citizens that don’t want the government to take over every aspect of their lives.

    Regardless of the outcome, the integrity of CBS, et.al. has long since been at the bottom of the dumpster.

  11. breschau says:

    “Perhaps the CBS poll that shows Barack Obama with a 14-point lead among likely voters (12 points when third-party candidates are included) is a modest outlier. But if so, John McCain has more and more outliers that he has to explain away these days. There are now no fewer than seven current national polls that show Obama with a double-digit advantage: Newsweek (+11), ABC/Post (+10), Democracy Corps (+10), Research 2000 (+10), Battleground (+13), Gallup (+10 using their Likely Voter II model) and now this CBS News poll.”

    http://www.fivethirtyeight.com/2008/10/todays-polls-1014.html

  12. breschau says:

    AJ:

    You have made several points about recent polls that point to the disparity in Democratic/Republican representation. I have a a couple of simple questions:

    Do you honestly think that the party affiliation in the country amongst those two is 50/50?

    If so — why? Can you give one piece of evidence that supports that assertion?

    If not, what do you think the current distribution is?

    And if not, what affect should the difference in your distribution vs. the distribution listed in the polls (actual and weighted) have on the reported numbers? Would it change any of the polls from “pro-Obama” to “pro-McCain”?

  13. crosspatch says:

    From October 7, 2004 poll by the AP.

    Among 944 likely voters, the Kerry-Edwards ticket led Bush-Cheney 50 percent to 46 percent. The Oct. 4-6 survey had a margin of error of plus or minus 3 percentage points.

    Not much different than what Rasmussen is showing right now.

    And Rasmussen’s tracking poll on October 4, 2004 had Kerry ahead 48.4% to Bush 46.4% .

  14. clintsf says:

    breschau-

    A poll is only as good as its assumptions.

    It’s certainly true that if Democrats show up at the polls in much greater numbers than Republicans do, Obama will win the election.

    But none of the pollsters are publishing the details of why they think that Democratic voters will have such a huge turnout advantage (more than twice the greatest turnout differential in the last 28 years).

    Could it be that ACORN’s 1.3 Million new registrations of Mickey Mouse and Tony Romo are factored into that assumption?

  15. Birdalone says:

    not sure any poll taken in the past two weeks is based on anything except panic.

    wonder what the polls would look like if Obama’s core supporters ever found out that Biden:

    “led the charge in the Senate for passage of the 1986 Anti-Drug Abuse Act, which — among its numerous notorious provisions — re-established mandatory minimum sentencing for drug crimes, expanded the use of federal asset forfeiture laws, and established the racially biased 100-to-1 sentencing disparity for the possession of crack versus powder cocaine. …

    Biden is also a staunch supporter of U.S. anti-drug efforts abroad, such as Plan Columbia and Plan Afghanistan, and has even espoused for the use of mycoherbicides such as Fusarium oxysporum — a genetically engineered fungal plant killer — in illicit crop eradication efforts. (Fortunately for the planet, more rational minds — at the U.S. Central Intelligence Agency and the U.S. Drug Enforcement Administration, of all places — nixed the idea, deciding that the deliberate spread of such toxic pathogens would be unsafe for the environment.)

    In recent months, Biden has called for a nationwide smoking ban, …”

    source: http://www.greenchange.org/article.php?id=3200

    Nationwide smoking ban? That’s 23% of voters! Someone should email NC and VA.

  16. ExposeFannyNFreddyNow says:

    If the polls are correct, and for the sake of argument the spread is +/-10%, then why so much anxiety here from the liberal left? Why would anyone from the left be wasting their breath here if this is such a shoo-in?

    Think about it.

    Teams tied top of the 5th, final game of a tied World Series, what’s the feeling, what’s the thinking? What are the questions?

    Now try bottom of the 9th, one side up 5 runs, would the winning team be on the mound debating the scoreboard with the other team?

    They say “the proof is in the puddin'” and there’s a whole lot of puddin’ here tonight, folks.

  17. The Macker says:

    Bresch,

    Based on the historical track record for the last ten years in terms of self-identified party affiliation from actual exit polls, the distributions are as follows: 38.4% Democrat, 35.8% Republican, 26.0% Independent, with only 2.6% point advantage for Democrat. And if we specifically use the exit polls from the last four presidential elections, the results are as follows: 37% Democrat, 35% Republican, 28% Independent. – Elwin Tobing at American Thinker

  18. AJStrata says:

    Folks, I noted in an addendum that there is no way only the indies moved and nothing showed in the Rep numbers.

    If there was a shift of one third of the indies, one would expect to see at least 10% of the GOP (moderates) go with them.

    This is a bogus poll for gullible friends Breschau and Conman to swoon over!

  19. archtop says:

    “And if we specifically use the exit polls from the last four presidential elections, the results are as follows: 37% Democrat, 35% Republican, 28% Independent.”

    This is exactly what makes these national “polls” ridiculous. Their internals nearly always show many more democrats sampled than republicans (often more than 10 points!!). So these polls are not a true sample of real America but merely a cartoon of a fake America contrived by the pollsters.

    As I’ve said before, John McCain simply needs to take the majority of the independents in the battleground states and he wins…

  20. sbd says:

    Apparently the AP is already prepared with the results of the election and has posted a Trial Run!!

    The following Associated Press “test” election results, marked “NOT FOR PUBLICATION”, were, presumably accidentially, released on the International Herald Tribune website at

    http://www.iht.com/articles/ap/2008/09/30/america/Elected-All.php .

    The IHT is owned by the New York Times Company and is in practice the “international” version of the New York Times.

    The list likely represents a test of AP’s election coverage system, with its predicted headlines having been written well in advance. The document was obtained by Wikileaks staff after a tip off from a reader.