I received the following timely email recently:
dear abby (hoffman)
i noticed the harris poll has kerry winning? i live in a republican stronghold but i'm seeing more and more kerry signs pop up. various other things indicate that folks are fed up with bush yet the media (corp.controlled) and major polls have bush winning. what gives?
signed, shocked and awed in illinois
Dear saaii,
The Gallup polls are full of shit and should be ignored, that’s what gives. I decided to lead off with this concise, easily understood answer, because whenever one delves into the sordid world of polling and statistics, or even mentions the word “statistics,” one is guaranteed to lose one’s audience faster than Pee Wee Herman. I’m going to avoid too much detail, because it’s not really necessary to understand all the ins and outs in order to understand what’s really going on.
If we take a quick look into the recent past (and it must be quick because history is almost as deadly as statistics), we get a strong indication that Gallup is no longer the crack organization we were always led to believe it was.
On October 27, 2000, just days before the last presidential election, Gallup released another seriously flawed poll. Again, this one was paid for by CNN and USA Today. It showed Bush leading Gore by 13%, with Bush at 52% and Gore at 39%. As we know, Gore actually won the popular vote, as well as the election, only to have it stolen by the forces of evil in the form of moss-backed, reprobate Florida officials and their criminal Supreme Court Justice duck hunting partners.
Both Time and Gallup have released two polls each since the end of this year’s Republican Convention that have been notable because of their outsized margins favoring Bush (from 8 to 13%).
The average of 17 other polls released during the same period showed Bush with a less than 3% lead. But the media trumpet the outliers. I'm not aware of the Harris poll showing Kerry ahead, but the Investors Business Daily, in conjunction with the Christian Science Monitor, just released a poll that even shows Kerry with a 1% lead.
Setting aside the question of outright fraud, why is Gallup getting such bad numbers? Gallup's methodology is based on the idea that Republican turnout on Election Day is likely to exceed Democratic turnout by six to eight percentage points. Yet exit polls from the last two Presidential elections show that Democratic turnout exceeded Republican turnout by four to five points. Just correcting Gallup’s numbers for their erroneous turnout assumption pretty much erases Bush’s imaginary lead.
Looking more closely at the last two Gallup polls reveals the possibility of a 10% swing in Kerry’s favor from Sept 15 to Sept 26. The analysis for this is appended at the end -- it’s just more fun with statistics that isn’t really central to the issue at hand.
The prospect of bias on the part of Gallup and CNN is very real. George Gallup Jr now directs Gallup’s non-profit research center. He is a devout evangelical Christian who has been quoted as calling his polling "a kind of ministry." And a few months ago, he said "
the most profound purpose of polls is to see how people are responding to God." Moveon.org just covered much of this with an ad placed in the NY Times:
Why Does America's Top Pollster Keep Getting It Wrong?Moveon.org’s charges reflect badly on CNN, who co-sponsored the poll.
Judy Woodruff did her best anorexic blond anchor “journalistic” investigation imitation into the charges the other day. As we can see by the Columbia Journalism Review's dissection of CNN’s coverage, it was basically a PR hack job:
Often, CNN covers contentious issues like this with sound bites from both sides, treating both positions roughly equally. But not this time. After all, a blow to Gallup's reputation as a reliable polling service is also a blow to CNN. So, on the network's "Inside Politics" this afternoon, it dealt with the issue this way:
Anchor Judy Woodruff began by briefly outlining MoveOn's complaint: "Recent polls have shown George W. Bush leading John Kerry and MoveOn.org claims Gallup's polling techniques exaggerate Republican support." Woodruff then gave Gallup editor-in-chief Frank Newport almost three minutes to respond, uninterrupted, to the charges. Naturally, Newport defended Gallup's methodology, but essentially asked viewers to take it on faith that he knows what he's doing.
End of segment.
With that nifty sign-off, CNN implicitly confirmed a criticism of itself that was leveled in the MoveOn ad: the charge that CNN winds up "acting as unquestioning promotional partners, rather than as critical journalists." For this was not the journalism of a disinterested party with no ax to grind. This was PR.
All of this points out the importance of understanding any bias in sampling that may be at work, even our own. Although you live in a "republican stronghold" which has Kerry signs sprouting like mushrooms, projecting local results onto the national level is fraught with difficultly. You’re in Illinois, a state in which Republicans are having major difficulties in general, and which is suffering under the ill-conceived candidacy of Alan Keyes. Democrats can expect to do significantly better there this year. Also, the area you live in, while typically Republican, isn’t the kind of place where one sees a lot of Confederate flags flying -- I’d think they’re more into wife-swapping than Bible thumping in your neighborhood. Although I also believe that
Democrats will do much better than the polls and pundits are suggesting, it's good to be aware of the self-reinforcing nature of much of the information we typically seek out. Watch Fox News for the next week and let me know what you think about Kerry's election prospects.
The point of all this puffing up of Bush's poll numbers at the expense of Kerry is to suppress Kerry's turnout by demoralizing Democrats and Democratic-leaning independents. If we believe that Kerry can't win, then come election day we'll be in a fairly negative frame of mind. It might lead some of us to decide to skip voting because of the inconvenience since it doesn't matter anyway -- Kerry's so far behind he can't possibly win! I don't believe it works to the same extent for the party which thinks it has the election handily won, and especially not for certain Republicans. There is a large segment of Republican voters who vote Republican because they have been deceived to see the party as one that projects strength, and because those voters want to be on the "winning team." (A close friend explained to me that wanting to be on the winning team is why he votes Republican.) Voting for the winner allows participation in the victory, and so tends to overcome those last minute excuses to skip voting.
A final note. On Sept 22, I blogged a quote from Mark Shields which bears repeating in this context:
One of the smartest Republican professionals I know in Ohio confided that he feared the GOP "needs a 5-point lead in the polls heading into Election Day" to counter what he sees as "the Democrats' intensity" and organizational commitment in the Buckeye state.
Democrats: Vote early and vote often!
Sincerely,
Thurston Howell IV
Some fun with statistics:
I don’t really believe that a 10 point swing has occurred in Kerry’s favor, but the argument can be made from Gallup’s own data. The Sept 13 - 15 poll that gave Bush a 13% advantage over Kerry had 7% more Republicans than Democrats in the polling sample (40% Republicans to 33% Democrats). The more recent Sept 24-26 poll showing Bush with an 8% advantage over Kerry had 12% more Republicans than Democrats in the poll sample (43% to 31%). So even though 5% more Republicans were sampled in the latest poll (going from a 7% advantage to a 12% advantage), Bush's supposed lead shrunk by 5% relative to the previous Gallup poll (a 13% lead dropping to an 8% lead).
Printer-friendly format
Email this thread to a friend
Bookmark this thread
(436 posts)
Send PM |
Profile |
Ignore
