Zogby/WSJ Polls (Senate)

Wall Street Journal ^ | 10/19/06 | John Zogby

[Ravi]

Republicans edged ahead in one tightly contested senate race but while the latest Battleground States Poll puts the party on track to retain control of the chamber, several tight races leave the Election night outcome uncertain.

[truthandlife]

FLASHBACK: Zogby Debunked: 2004 predictions, worst ever

[USS Alaska]

Either the tide is starting to turn again, or the polsters have to start getting honest as the election approaches.

An honest pole.

[Just A Nobody]

;*) I concur

[rwfromkansas]

I am to the point that if we can at least get the Senate, I will be at least okay wtih the election.

Why? Perhaps if we are the opposition party in the House, it will fire us up to actually DO something in Congress.

But, the Senate is vital...we need it in case of a SCOTUS retirement or Dem actions in the House to try to investigate people.

[Rokke]

"Perhaps if we are the opposition party in the House, it will fire us up to actually DO something in Congress. "

The minority doesn't ever "DO" anything other than obstruct things from getting done.

[joesbucks]

OK. So now we trust Zogby?

Only when it's good news to us. Then he's a genious, has a strong verifiable and repeatable polling technique and worthy of our respect. Otherwise, when his numbers go against us, he's MSM scum.

It's the conservative world view.

[phillyfanatic]

For two straight weeks, I want to see Zog, Rass, and Battleground on the same page. If the Pubs are up in the Senate races and will hold the Senate by , say Nov. 1, I will believe these polls. As for the House, it is now up to 27 seats supposedly the Pubs will lose. I want to see those races all polled by the 3 pollsters above and by Nov. 1 want to see how they play out. It may be the Dems win the House but if they do, it will be by a few seats not a landslide. If they win the Senate, it will be by one or two seats. And if the Gop holds, it will be 51-49 or 52-48. Now, we need all Pub purists to come back to the Party and vote for the Pubs in Nov. instead of cutting and running and staying at home.

[LS]

I think that's about right. Again, our internal polling of Republicans---cause if we get the Rs to vote, Ken wins---is showing a real rapid coalescing of them around Ken. Between that and the real turnout numbers, Blackwell will win a squeaker. And boy, will the Dems be angry.

BTW, I've not seen you post before, but this was really impressive. Excellent research and analysis.

[LS]

But in 1996, every one of those "reliable" polls were off, many by far more than the MOE, and every last one of them was off in Clinton's direction. Gerry Wasserman of Purdue did a statistical study that found that such a result being arrived at randomly was 240,000:1.

[LS]

We'll celebrate after the election!

[DontBelieveAugPolls]

i know those polls in 1996 were off ,but they did predict correctly the winner.

if you are correct, we are about to have the most surprising election since 1948.

Even with generic polls showing the GOP behind in2002 nad 2004, you were hard pressed to find any GOP incumbents in trouble...in 2006 the list of GOP incumbents in trouble is probaly in the range of 20+

[LS]

Well, Zogby, for one, was WAY off in his "picks" in 2002, and both Cook and Sabato had far worse predictive records in both 2002 and 2004 than I did. I only missed two competitive races in two campaigns: Thune in 02 (500 votes) and Coors in 04 (pretty close).

But it's still irrelevant. To say that a) the polls have all---EVERY ONE---have been badly off before, and b) that we have evidence from THIS CYCLE that some polls have already been found to be oversampling Dems heavily, equals c) the polls lie . . . most likely, all of them.

[KMAJ2]

[I don't say the WSJ is liberal...]

You need to separate the WSJ news operation from their editorial board. The WSJ news operation is decidedly leftist, while the editorial board is conservative. Zogby is aligned with the WSJ news division, not the editorial side. This is a common misconception that the WSJ is a conservative news outlet, it is not. In fact, a 2004 study, "A Measure of Media Bias," by Tim Groseclose of the University of California, Los Angeles and Jeff Milyo of the University of Missouri, stated that:

"One surprise is the Wall Street Journal, which we find as the most liberal of all 20 news outlets [studied]. We should first remind readers that this estimate (as well as all other newspaper estimates) refers only to the news of the Wall Street Journal; we omitted all data that came from its editorial page. If we included data from the editorial page, surely it would appear more conservative. Second, some anecdotal evidence agrees with our result. For instance, Reed Irvine and Cliff Kincaid (2001) note that "The Journal has had a long-standing separation between its conservative editorial pages and its liberal news pages." Paul Sperry, in an article titled the "Myth of the Conservative Wall Street Journal," notes that the news division of the Journal sometimes calls the editorial division "Nazis." "Fact is," Sperry writes, "the Journal's news and editorial departments are as politically polarized as North and South Korea."

http://www.polisci.ucla.edu/faculty/groseclose/Media.Bias.8.htm

So do not confuse the conservative WSJ editorial division with its left wing news division, which the study found as the most left wing of ALL news outlets studied.

[Recovering_Democrat]

Wow. I was unaware of that study. Thanks for the info; I'll consider it very carefully!!! :)

RD

[KMAJ2]

The same reasoning applies to the WSJ alliance with NBC's polling. Once again, it is the WSJ news department aligning with NBC to do the polling, not the WSJ editorial department. I have heard, on more than one ocassion, the claim because of the affiliation with the 'conservative' WSJ, NBC's polling results can be trusted.

The most accurate polls in 2004 were Rasmussen, Mason-Dixon and SurveyUSA, Rasmussen was # 1 in both sum and spread methods of evaluation.

Let's Go to the Audiotape
Who nailed the election results? Automated pollsters.
By David Kenner and William Saletan
Posted Thursday, Dec. 9, 2004

http://www.slate.com/id/2110860/

[jmj3jude]

I have read many of your posts, and they bring great comfort. I have a sister and brother who live in Ohio, they say the local news is very depressing. I just heard on Shawn Hannity show and Bill Cunningham said that the local GOP party has given up on Ohio. That sound really depressing and I hope that it is just wrong. What I don't understand is that the dem's running are so liberal and Strickland with all the crap with the guy in his campaign exposing himself would get a lot of attention and cost him the election, it should. DeWine I am worried about, maybe you could shed some light on that race. Thank you!

[LS]

I get snippets locally, and, of course, keep my eyes open. I'm in Montgomery Co., a swing county that Bush lost by 3,000 votes. I'm hearing Blackwell is now ahead here, and pulling 40% of the black vote statewide. I take the latter somewhat with a grain of salt, as it seems more blacks are "always" voting GOP until the actual lever is pulled. But, heck, if Ken gets 30%, Strickland is in big, big trouble. NO ONE on the ground---NO ONE---believes the Strickland/Blackwell numbers. If the national GOP has given up, it's news to us, because Blackwell's campaign only two weeks ago had the national people come in and bring workers and money.

I don't know about DeWine because I'm not working for his campaign. Everywhere I go, I hear it's "tough," because of his ridiculously stupid votes. However, the bottom line in OH is this: we have the numbers. Bush won OH by 115,000 votes after losing Montgomery. I'm seeing very little "anti" feeling toward the GOP, but callers have told me that many people are saying, "I will vote a straight GOP ticket, but will leave senator blank."

Will they do that? Who knows. My priority is, a) GET THEM TO THE POLLS. If we do that, we'll have a real, real good chance for b) get them to vote for ALL the GOP. If those two things happen, Republicans will win OH, period.

Last point: in all our walking, dropping lit, etc., from both personal observation and reports from others---and realize, we have people out six days a week in two shifts a day!!!---that there are no Dems out. NONE. ZERO. I asked the Mont. Co. Blackwell chairman about this and she said, "They're just like they were in 2004," which is to say, inefficient, incompetent, and incapable.

[Dr. Free Market]

The problem I'm finding is that some of the pollsters are weighting their results to the 2004 exits, which we know were way off instead of previous midterm elections. That's why I prefer the last U of Akron Bliss Poll for Ohio that weighted to both 2004 and 1998 (open Ohio gubernatorial election). The problem with their subsample of "likely voters", however, was that only 80% considered economic issues important. That 80% is probably the true midterm "likely voter" in this Ohio election. The 20% that didn't is in the range of the usual drop off between presidential and midterm elections (10-23%) in Ohio. Granted these are assumptions and increase the MOE, but this leads one to assume that their point statistic (the number that gets widely reported) is actually closer to the upper range of the MOE (+/- 4.5) instead of the middle. This makes more sense and is in line with what is being reported by the campaigns.

Most polls allow respondents to self-select as "likely voters" without asking confirmatory questions of some type. Rasmussen, for example, asks if you have voted in the last 4 general elections-- better than nothing-- but some people interpret that as presidential only, forgetting about midterms, but some include primaries.

Unless you correctly define "likely voters", appropriately weight the respondents, and ensure that respondents are similar to non-respondents (Zogby does), you're going to get skewed results like we're seeing. No poll is perfect, but some are much worse than others.

Pay attention to the methodology of polls when they come out. If they don't report the 3 things above and the internals (breakdown of the characteristics of the respondents), don't believe them. You'd be surprised how many of these "professional polls" that get widely reported don't do these things-- mostly because of cost and time constraints. The internal polls of the campaigns are done more correctly and follow trends, but they don't get reported for strategic reasons-- usually.

[Norman Bates]

sitetest, I would call it a dead heat which is bad news for Cardin.

[KMAJ2]

All polls should be taken with a grain of salt, nevermind the weighting, but the simple choice of words in a question can skew results. Poll reporting is, in my opinion, lazy journalism. What I find more interesting is that the polls cited most often by the MSM are not those that were the most accurate in the past election. That does not mean those polls will again be the most accurate in this election, thus the reason all should be taken with a grain of salt.

The one thing that none of the polls take into effect, is the effectiveness of the get out the vote efforts of each party. In the close races, that will tell the story of who won and who lost. The report card on the pollsters after this election should be interesting with the wide disparity in figures being reported.