Zogby released a poll yesterday showing Hillary Clinton losing potential match ups with five Republican nominees (even Mike Huckabee), a dramatic shift from a few months ago when Hillary was favored over all Republicans in Zogby polls. In the same match ups, Obama comes out on top and Edwards either ties or wins against most Republicans.
So what happened? The Clinton camp questions the methodology and claims the poll is inaccurate because it is Zogby's first online poll. Granted, the results probably shouldn't be compared with previous polls that were conducted via landline telephones, but which of the two methods is more accurate?
There is some validity to the argument that Internet polls exclude a (small) portion of the population without access—perhaps older Americans who are more likely to vote. But traditional telephone polling overlooks a large segment of the population who no longer has a landline and instead relies solely on cellphone service.
And compare the sample sizes: The telephone polls had samples of around 1,000, whereas this latest online survey included more than 9,000 likely voters and had a margin error of +/- 1.0 percent.
My endorsement goes to the online survey. Hillary is beginning to slide, and rather than acknowledge that, her campaign wants to cast doubt polling methodology. Her game plan is to exude a Bush-like confidence (i.e., arrogance) regardless of what's happening in the polls. Consider her interview last night with Katie Couric:
COURIC: If [the nominee] is not you, how disappointed will you be?CLINTON: “Well, it will be me...”
COURIC: I know that you're confident it's going to be you, but there is a possibility it won't be. And clearly you have considered that possibility.
CLINTON: “No I haven't.”
COURIC: So you never even consider the possibility?
CLINTON: “I don't.”
She had better consider the possibility soon, because the latest poll reflects what has been a common assumption for many: Clinton is too polarizing to draw support from disenchanted Republicans (there are many) and independents and won't do well in a general election.
Like this post? Get updates via|
Comments (6)
Ug! Hillary's blind arrogance made me throw up a little in my mouth.
Posted by James Sardoco | November 27, 2007 12:36 PM
Neither methodology is safe. I wouldn't place a bet based on telephone surveys nor internet surveys.
Sure, landline usage is in decline, but the numbers were probably more accurately collected. E-mail address samples are very hard to collect and are often self-selecting, both of which are problems for survey research. The nonresponse incidence in E-mail surveys is also higher, partly due to spam problems and partly due to people changing E-mail addresses frequently.
If I *HAD* to choose a methodology, I'd go with the landline results. Though the margin of errors are different because of the sample sizes, the method of sample selection is probably more secure with landlines.
Posted by Jofeshenry | November 27, 2007 2:25 PM
I agree that both methodologies are questionable. However, the Clinton campaign is using the shaky online methodology to dismiss the trend entirely, and I do believe there are valid results in the latest poll. She seems to be losing momentum, and not just in Iowa.
Posted by Elyas | November 27, 2007 2:45 PM
Please read some statistics book before claiming that internet polling gives a better sample.
Bigger sample ,maybe, but better, I will argue not.
Internet being self selective polling, I do not give any value to these numbers anyways.
Also, I do not give any value to the biased media poll results.
Enjoy.
Posted by Pt | November 27, 2007 2:52 PM
Probably good information but this "world news feed" box covers up almost half your page, obscuring the text........................................
Posted by RLM | November 27, 2007 6:47 PM
funny. i didn't even notice the "world news feed" until you pointed it out. i guess some people are "readers" and some people read by "lookin' at the pictures".
Posted by eric | November 27, 2007 9:29 PM