Why Zogby's polls skew toward McCain.

[pnwmom]

They poll "likely voters" rather than "registered voters" -- leaving out newly registered voters, who are more likely to vote Democratic.

Unlike PEW, they don't poll cell phone users (who are more likely to vote Democratic)

They weight the sample by PARTY. In other words, they have arbitrarily decided that there SHOULD be a certain fraction of Republicans in the sample. If fewer people than that actually claim to be Republicans, then they weight the responses so that a Republican response COUNTS MORE than a Democratic response.

Other more reputable firms weight by gender and other characteristics (to make sure they have a representative sample), but not by political party, because party affiliation is fluid, not fixed. Zogby pretends that party affiliation is known and doesn't change. In other words, if "not enough" Republicans turn up in a sample, they assume that the poll has a sampling error that needs to be corrected -- rather than that the Republican party has its members deserting in droves.

http://www.zogby.com/methodology/readmeth.dbm?ID=1341

Released: October 08, 2008

Survey Methodology Reuters/CSPAN Zogby National Tracking Likely Voters 10/4/08 thru 10/7/08

Zogby International was commissioned by Reuters and C-Span to conduct a telephone survey of likely voters.

The sample is 1220 likely voters interviews with approximately 39 questions asked. Samples are randomly drawn from telephone cd’s of national listed sample. Zogby International surveys employ sampling strategies in which selection probabilities are proportional to population size within area codes and exchanges. Up to six calls are made to reach a sampled phone number. Cooperation rates are calculated using one of AAPOR’s approved methodologies and are comparable to other professional public-opinion surveys conducted using similar sampling strategies. Weighting by region, party, age, race, religion, gender is used to adjust for non-response. The margin of error is +/- 2.9 percentage points. Margins of error are higher in sub-groups.

[Ozma]

1. In other words, Zogby LIES to make money......like most Rethugs

I figured as much.

Last time I talked to a REthug was when he wanted to sell me something I didn't need.

[regnaD kciN]

2. Actually...

...many tracking polls are based on likely voters rather than registered ones. Although this year may be an exception, LV polling is generally more accurate than RV.

And virtually all tracking polls, except for Gallup, weight by party as well. The question is: what percentages is Zogby using? Diagero/Hotline went from a 6-point Obama lead to a 2-point Obama lead yesterday...but also changed their weightings so that Democrats only had a 2% turnout advantage rather than the 5% they had been using earlier. If you go back to their old weighting, the result barely changes from the day before, instead of being a big Obama drop.

[Ozma]

3. Yes, I know that about LIKELY.....

Well, those that took the trouble to register in the last 3 months did NOT do it to elect McCain........


And they are NOT EVER going to forget to vote!!!!!!!!!!!!!!!!!!!


18 to 28 year olds........WILL VOTE......even IF they never voted before..and are NOT considered "likely"




90% or more of them will VOTE!!!!!!!!!!! So they are "LIKELY voters"

[pnwmom]

5. But when a pollster determines "likely voters," what they mean is:

people who have reliably voted in the past.

So newly registered voters are underrepresented in this scenario.

[pnwmom]

6. delete (dupe) n/t

[pnwmom]

4. I'd like to see the data that says "virtually all" tracking polls

weight by party as well.

The National Council on Public Polls is a reliable source of information about polling.

http://www.ncpp.org/?q=node/39

"Bad Weighting: The most common bad weighting in political polls is weighting just the likely voters for the number of Democrats, Republicans and others. Party identification is correlated with voting, but what is missing are meaningful numbers of party members for the whole population. Without that, the weighting is a guessing game rather than good theory. Some pollsters use the numbers from an exit poll from a past election, but the numbers of people who consider themselves members of a party changes from month to month and year to year.

"Consider the change in party identification from the Pew Research Center polls throughout 1996. In the beginning of the year the Republican-Democrat split was 30%-30%. On Election Day it was 26%-36%. The number who considered themselves Republicans went down steadily the closer the survey was to the election. Using party identification to weight just the likely voters in a political poll is little better than a guessing game where the pollster is substituting his or her judgement for scientific method.

"Another type of bad weighting for elections polls occurs when the weighting characteristic has a low correlation with key variables such as voting. For example, in some elections there may be no relationship between age and voting. If the age-voting relationship is weak then weighting for age will make the sampling error larger. Sometimes there is a strong relationship with vote but a weak relationship with some other characteristic. That other characteristic will have a larger sampling error. One also should avoid weighting that creates a large discrepancy in the weights applied to different sample cases.

"The rules of weighting are simple. Always weight for the different probabilities of selecting the sample and be cautious about weighting for other things. "

[regnaD kciN]

7. It was covered on 538.com...

...a few weeks ago. Don't know the exact date, but I'm pretty sure it was after the Democratic Convention.

[pnwmom]

8. I've actually heard the opposite -- that most polls are weighted, but

NOT by party affiliation.

In any case, I think what the National Council on Polls has to say on this makes a lot of sense. We know that the numbers of people who identify as Republicans are dropping, so trying to weight a poll in this fashion is pure guesswork. The number of people who claimed to be Republicans in the exit polls in 2006 has little to do with the number who consider themselves to be Republican today.

[mucifer]

9. All I know is Zogby predicted a Kerry win. They were right.

Keep that in mind.

[Bonn1997]

13. Yeah, I don't get why they would have done a 180 in 4 years

[Joe Chi Minh]

15. Yes, weird. I remember posters remarking that they were the most honest

of the bunch. What they're doing now, is beyond ludicrous. Even for a polling company. But they wouldn't be the first outfit or person who's completely changed direction for the worse under this administration. It seems to be a feature of public life now that these totally anomalous 180 turns occur in people and organisations that had hitherto earned the public's respect.

[nickn777]

10. On the EVE of critical Democratic primaries.........

Ohio:
Zogby--TIE
Actual Results--Clinton +10.1

New Hampshire:
Zogby--Obama +13
Actual Results--Clinton +2.6

California:
Zogby--Obama +13
Actual Results--Clinton +9.6

[Bonn1997]

14. Actual results or *reported* results?

[Loge23]

11. In So. FL, a whole lot of Obama stickers

I don't know about all of these polls. All I know is that I am seeing many Obama bumper stickers in South FL.
Easily, 4 to 1 Obama on the unofficial bumper sticker poll on I95!

[GalleryGod]

12. Correct! Likely leaves out NEW Voters.

Well there you go again, Zogs!

[THUNDER HANDS]

16. he was probably shaken from 4 years ago

and decided to change things to reflect more Republicans in his polls.

[BlooInBloo]

17. They have nearly the worst track record in the business.

[Onlooker]

18. On the other hand, I wonder if they factor in the Bradley effect?

It concerns me that Zogby's predictions are relatively negative. After all, Zogby is Lebanese (Christian), calls himself a Democrat, and is active in Arab affairs. Given that, I'd have to believe that Zogby personally would rather Obama won, but he also may believe that racism is skewing the polling results. At any rate, if Obama is ahead in a Zogby poll, it probably means that even those of us who fear the Bradley effect is still around can breathe a little easier.

[CreekDog]

19. Because Zogby wants to be an outlier which gets his poll attention

and gets him media coverage. take a look at his polling over the long term, being the outlier is the norm for him and it always leads to his poll getting more coverage than the average poll --that's the one consistent theme in all of this.

he's pretty shameless.

[democrattotheend]

20. It seems like party weighting helps McCain

I know the GW/Battleground poll stopped weighting by party and Obama shot up when they did that. I wonder why they did weight by party...weighting by other demographics makes sense, because there are census statistics on the racial, age, gender and ethnic makeup of an area and it makes sense to ensure that the poll has a representative sample. But party ID is a moving target...it often changes with the presidential vote. When McCain shot up after the GOP convention, it coincided with an increase in GOP party ID and GOP gains in the Congressional ballot.