Zogby 2000

[Shira]

In 2000 he was better than most pollsters
http://www.zogby.com/news/ReadNews.dbm?ID=299

[Fmr. Gov. NickG]

Yeah, I think it was his record in 2002 that is the major source of embarassment.

[CARLHAYDEN]

There is a very simple reason for Zogby being good in 2000, and lousy in subsequent elections.

In 2000, the electorate was skewed Democrat as the result of a very effective activation effort by the DNC, the Gore campaign and its affiliatiates.

The funding for that effort went back as far as 1996, with initial research in 1996, and initial implementation in 1998 (when Democrats did better than historical models or expectations due to a Democrat skewed electorate due to an effective activation effort in several states).

The money and the organization ran out, and has not been rebuilt.

However, the Bush campaign learned from 2000 and put money and effort into their own activation effort, which should skew the electorate in a Republican direction this year.

Zogby is still assuming the electorate of 2004 will mirror the electorate of 2000.

[Shira]

Zogby is still assuming the electorate of 2004 will mirror the electorate of 2000.

I wonder how his "polling views" or his assumptions (if any) could change the numbers he gets over the phone in favor of Kerry.

[The Duke]

Zagby runs a weighted poll.  If his turnout assumptions are true, so is his poll.  Hence he is usually erratic, but sometimes brilliant.  Donna Brazille made his turnout assumptions true.  To make the '04 electorate look like the '00 electorate, and therefore make Zogby's assumptions true, the Democrats will have to get the saem phenomenal turnout as last time and the Republicans will have to do just as poorly as last time.  Won't happen.

[Shira]

As to turnout, my prediction is that it is going to be higher by 8% than last time.
The common view among the pundits is, that a high turnout works in favor of the Dems.
We'll se.

[MODU]

Zogby got lucky.  He has yet to come close this year.

[The Vorlon]

My problem with Zogby is he makes a ton of assumptions.. and then does not tell you what they are.

He also, imho, steps over the line between polling and prognositication.

For example, if say Charlie Cook said "The Polls show "X" to be ahead by 5%, however "Y" has a really good get out the vote effort so I think the race is actually very close" that is a valid and reasoned OPINION that presents his logic and people can make an informed judgement upon it.

If Zogby thought the same thing, he would just just "juice" turnout for "Y" in his polls and present the race as being tied as a FACT.

Two very different things.

[afleitch]

Zogby hits the nail on the head when it comes to trends...not actual results. When focusing in on on Iowa during the caucuses he saw the rise of Kerry and Edwards in the two weeks before the caucus itself. He also was the first (well non subscription at least) pollster to show Missouri as swinging towards Kerry, a state most pollsters ignored until July, and also reported on a closer race than expected in Tennessee, something the traditional pollsters only just realised this month. If you want your trends, look at Zogby, if you want figures, look elsewhere. And as for Gallup, theres something a bit wrong with them. Oh and try the British newsmagazine, The Economist for its American YouGov poll, another good trendsetter, even if the percentages themselves are skewed.

[muon2]

As to turnout, my prediction is that it is going to be higher by 8% than last time.
The common view among the pundits is, that a high turnout works in favor of the Dems.
We'll se.

It's hard to see where there would be such a large jump. One of the experts in voter turnout is Prof. Michael McDonald at GMU. He has a website with a wealth of analysis. One important technique he uses is to measure a turnout rate by subtracting the iinelgible non-citizens and felons from the voting age population, and also subtracting the overseas vote from the total. For the last few elections this reconciled turnout rate is:

1948 - 52.1%
1952 - 62.2%
1956 - 60.0%
1960 - 63.6%
1964 - 62.6%
1968 - 61.4%
1972 - 56.1%
1976 - 54.7%
1980 - 54.6%
1984 - 57.0%
1988 - 53.9%
1992 - 60.2%
1996 - 52.2%
2000 - 55.2%

From this you can see that since 1972 after the voting age was lowered to 18, the turnout exceeded 60% only once, and that was in a three-way race that brought out new voters.  A 5% increase would be consistent with that recent record, and more likely it will stay within 1-2% of the turnout from 2000.

[MODU]

That was when us Perot-Reformists attempted to make our stand . . . and failed.  hahaha

[The Vorlon]

In 2000 he was better than most pollsters
http://www.zogby.com/news/ReadNews.dbm?ID=299

Sorry... but you are just plain wrong.

Zogby has missed the margin of error in 20 of his last 45 polls going back to 2000.

In 2000 he missed MOE in 11 out of 31 polls.
In 2002 he missed MOE in 9 out of 14 polls.

These are dead cold facts.

Ignore the spin on his website.  He did 31 polls in 2000.. yes a few were dead on, and a few more close.

Look at the totality of his work.

If I had my cat pick a horse to bet on 45 times I am sure I'd get a few winners... Does this make my cat a racetrack betting expert...?

Just in case ANYBODY has forgotten how truly laughably bad Zogby has been in the last 4 years...

Link to Zogby's 2000 poll results

Link to Zogby's 2002 poll results

I wonder if somebody can FORCE these graphics to be attached to all Zogby polls - maybe the "truth in advertising" laws or  hazardous product labeling or something...?