How was the the Newsweek poll conducted? How do you interpret this data?

PR Newswire ^ | 10/03/2004 | Whitman

[timbuck2]

SAMPLE SIZE/MARGIN OF ERROR FOR DEBATE VIEWERS SUBGROUPS: 770 Debate viewers (those who say they watched at least some of the debate) (plus or minus 4.1)

369 Men (plus or minus 6) 401 Women (plus or minus 6)

265 Republicans (plus or minus 7) 274 Democrats (plus or minus 7) 215 Independents (plus or minus 8)

NOTE: Data ares weighted so that sample demographics match Census Current Population Survey parameters for gender, age, education, race and region. Sample sizes listed above are unweighted and should NOT be used to compute percentages.

So how were the percentages generated?

[timbuck2]

Here is the conclusion by Newsweek:

The debate erased the lead the Bush/Cheney ticket has held over Kerry/Edwards in the Newsweek Poll since the Republican convention. In a three-way trial heat including Ralph Nader/Peter Camejo, among registered voters Kerry/Edwards leads Bush/Cheney 47 percent v. 45 percent with 2 percent for Nader/Camejo. In a two-way heat, Kerry/Edwards leads 49 percent v. 46 percent for Bush/Cheney, the poll shows.

**Okay, how is this conclusion drawn? Is it based on only those who watched the debate or based on the entire sample regardless of who viewed the debate?

Can somebody help me understand these sample, which seemed skewed toward women, Dems and west coast reisdents? I am confused.

-T

[conservativecorner]

The Media by all signs are going to push the following line this week. "Kerry makes comeback in polls and closes gap on major issues." Matt makes the point below that there appears to be collusion between the "paper of record" and the DNC/Kerry campaign. The President needs to campaign and debate as if he really is down by 5 points, and continue to hit John Kerry with his inconsistencies and nonexistent "plans". This as the saying goes is, "where the rubber hits the road", and we’ve got to keep the pressure on and get the word out however and wherever we can. 4 more years will be reality in just 30 days. ****Newsweek is the first out with the Kerry Comeback poll, here are a few sites looking at the wacky numbers.
http://powerlineblog.com/archives/008037.php


http://www.redstate.org/story/2004/10/2/182143/492


http://politicalvicesquad.blogspot.com/2004/10/liberal-media-cognitive-

dissonance.html

http://instapundit.com/archives/018191.php

[conservativecorner]

So, Yahoo! News now has an article titled "Bush Raps Kerry, Slips in Polls." It's based on a recent survey by Newsweek. Unfortunately, this story is a perfect example of why you should treat all media reporting on polls with skepticism. As Powerline notes:

"Newsweek's most recent poll included 345 Republicans, 364 Democrats and 278 independents. This compares to Newsweek's published data for their most recent prior poll, which showed President Bush with a comfortable lead: 391 Republicans, 300 Democrats and 270 independents. Yes, if you drop 46 Republicans and add 64 Democrats, you will get considerably better results for the Democratic nominee." (Hat Tip: The Pink Flamingo Bar & Grill)

The stuff above is called "screening" or weighting" the sample, and it's just one example of polling methodologies you ought to be wary of. Jay Caruso has more on this particular poll.

Other current polls don't seem to square with Newsweek's latest, just as other polls didn't square with Gallup's accounts of a 13% Bush lead a couple weeks ago. Usually, this is just sloppiness rather than bias - but these sorts of gambits can and will be used deliberately by either side.

If polls matter to you, therefore, get used to asking for internals, questioning the questions, and comparing different polls. Caveat elector.

http://story.news.yahoo.com/news?tmpl=story&cid=514&e=2&u=/ap/20041003/ap_on_el_pr/bush

http://www.papadoc.net/PinkFlamingoBar.html

http://jaycost.blogspot.com/2004/10/problems-with-102-newsweek-poll.html


http://smythesworld.blogspot.com/2004/09/another-wacky-poll-from-gallup-this.html

[conservativecorner]

Look at the makeup of Republicans vs. Democrats in Newsweek’s poll from September 11, 2004: NEWSWEEK POLL: Campaign 2004.

391 Republicans (plus or minus 6) 300 Democrats (plus or minus 7) 270 Independents (plus or minus 7)

Compare against the same data from the new poll, which Newsweek is using to claim that Bush’s poll lead has “evaporated:” NEWSWEEK POLL: First Presidential Debate.

345 Republicans (plus or minus 6) 364 Democrats (plus or minus 6) 278 Independents (plus or minus 7)

Did Newsweek choose a lower percentage of Republicans for the first debate to set up Kerry’s “comeback,” or did they stack the deck with more Democrats in the second poll?

(Hat tips to all who emailed about this.)

UPDATE at 10/2/04 6:54:49 pm:

The loons at Daily “Screw Them” Kos are watching this topic: LGFers moan about Newsweek poll.

[cripplecreek]

Does no one remember the DNC e-mail that circulated before the debate. It urged democrats to hurry and vote for john kerry as the winner. It was posted god knows how many times in this forum.

[Publius6961]

Wrong questions.
Ask me if I care.

Polls don't win elections. Votes do!

[Barlowmaker]

215 "independents"?

LMAO.

Like the AFSCME operatives who call C-SPAN with anti-Bush diatribes every morning on the "independent" line?

The country is NOT comprised of one-third "independents", unless the definition of independent is a "lying Democrat". There's your Newsweek anomoly.

[timbuck2]

Unfortunately, bogus polls (I am not sure if the Newsweek poll is bogus because nobody can seem to answer the wieghting issue) can tilt an election one way or another by motivating the dispirted side. So, polls DO matter, obviously.

-T

[fhlh]

LOL.... all of us loved the polls before the debate

[litany_of_lies]

The sampling IS skewed towards West Coast Residents and women compared to a similar survey done 3 weeks ago.

HERE is the Little Green Footballs post from last night, based on input from many, including yours truly:

---

http://littlegreenfootballs.com/weblog/?entry=12951_Newsweek_Poll-_Stacked#comments

Look at the makeup of Republicans vs. Democrats in Newsweek’s poll from September 11, 2004: NEWSWEEK POLL: Campaign 2004.

391 Republicans (plus or minus 6)
300 Democrats (plus or minus 7)
270 Independents (plus or minus 7)


Compare against the same data from the new poll, which Newsweek is using to claim that Bush’s poll lead has “evaporated:” NEWSWEEK POLL: First Presidential Debate.

345 Republicans (plus or minus 6)
364 Democrats (plus or minus 6)
278 Independents (plus or minus 7)


Did Newsweek choose a lower percentage of Republicans for the first debate to set up Kerry’s “comeback,” or did they stack the deck with more Democrats in the second poll?

(Hat tips to all who emailed about this.)

UPDATE at 10/2/04 6:54:49 pm:

---

In addition, here's what I said last night about the poll:

---

http://www.freerepublic.com/focus/f-bloggers/1233733/posts?page=3#3

But here's the biggest point:

In other words, they decreased Republican sampling by 5 percentage points and increased Democratic sampling by 6 full percentage points. Furthermore, this "poll" strictly was limited to the "Pacific and Mountain time zones," and was taken immediately after Thursday's debate. In other words, registered voters from the following states were excluded: Texas, Florida, Ohio, Pennsylvania, Missouri, and the entire old south.

So what they TRIED to do is get a poll based PURELY on post-debate reaction (they couldn't do this in the east, Midwest and south because it was too late to call). They were hoping that the reaction to Kerry's debating "skills" would cause a Kerry bounce.

When you correct for the flawed sample, it's clear they were WRONG.

Bush led by 6 in their September 11 poll. With an 11-point swing in sample of Dems and GOP, you would have expected Kerry to come out with a 5-point lead (same 11-point swing) if in essence no one had changed their minds--BUT IT'S ONLY 2. KERRY LOST GROUND!

A sample weighted the same as September 11 would have shown Bush lengthening his lead from +6 to +8 or +9. I happen to think both samples didn't represent the true DEM-GOP makeup of the country, which IMHO is roughly 37% GOP, 33% DEM, and 30% INDY. With that weighting laid over Newsweek's results, Bush's lead would be about +5.

Conclusion: BUSH came out of the debate with the momentum, which will only grow when the "global test" commercials and the next round of Swift ads hit the air.

---

And here's a final point: Since this was a Thursday night poll, not only did they not call the East, South, or Midwest at all (too late in Central and Eastern Time Zones), they only had ONE hour to call the more conservative Mountain states, and TWO hours to call CA, WA, and OR.

On the relatively safe assumption that the average Democrat in those three states is more liberal than the ones in the Mountain states, that further skews the sample.

AND DESPITE ALL OF THIS BOOK-COOKING (as noted in the text of my FR post last night, when the sample makeup is corrected for, BUSH IS BETTER OFF NOW THAN HE WAS THAN IN THE SEPT. 11 POLL

[Owen]

No.

No pollsters other than Rasmussen "choose" the partisan mix for their calling. They can't. The registration lists have no phone numbers on them.

The most they could do is call, ask, get the response and toss the results that are in excess of their desired mix.

But they don't do this because it would not make sense. Here's why. First, let's discard the conspiracy option. Let's presume they are trying to take a real measurement and not get a desired result.

If they want a real measurement they will sample at random and get the partisan mix that they get. It is not reasonable to conclude that people are changing partisan affiliation every 3-4 weeks (note heavy dem sampling in polls prior to late August, then heavy GOP Aug/Sept, now heavy Dem again). So the question is why do random phone calls yield such wide variances in partisan mix result.

For this Newsweek poll, the answer seems to be that they called only the west coast on Thursday night because the debate ended too late to call the midwest and east. That hugely emphasizes CA, OR and WA, all Dem strongholds.

They also called only Saturday morning because the poll was out mid-day yesterday. Thus, the bulk of calls were Friday night -- and how this skews is unpredictable.

Oddly, Rasmussen looks very good out of all this. He applies a desired partisan mix and does not vary it. His absolute numbers may not be valid because no one knows what mix will vote, but by holding it constant his trend data is the very best we have to go with. It has shown little change all throughout the campaign.

[timbuck2]

So, are these numbers

369 Men (plus or minus 6)
401 Women (plus or minus 6)
265 Republicans (plus or minus 7)
274 Democrats (plus or minus 7)
215 Independents (plus or minus 8)

supposedly in line with census numbers? Or are these numbers given a weighting so that they do reflec tthe census data? Somebody must know the answer to that question...

-T

[Allegra]

News-weak polls had Carter ahead of Reagan consistently through the election season in 1980.

'Nuff said.

[RedRepublic]

Hi guys,
Stop reading the bias media polls. Even before the debate took place we all knew they would push Kerry. Why bother reading them. Those polls will make us depressed. I think we all should just be optimistic and work harder to show our support to the President and to mobilize our base as well as to attract the "undecided" voters.
Look at the substance of the debate! Dont look at the style Kerry performed on the debate. Point out the "global test" doctrine. Tell the "undecided" voters to listen carefully Kerry's self-contradicting words. Hey, I haven't heard any clean "winning plan" from Kerry, except the word "I can do better" that he kept repeating.

[Publius6961]

So, polls DO matter, obviously.

To the dispirited, obviously. And the ignorant and the feeble-minded. They can all join DU and look silly for all I care.

I'd rather just make sure that committed voters with a brain bother to vote.

[Barlowmaker]

I'm with ya!

[Pali]

Step away from the poll....step away from the poll...

I'm going to assume the Newsweek poll is correct and focus my energy on getting President Bush re-elected. I don't think Pres. Bush had his Game Face on but I truly believe next time he'll do better. We all know he won on substance but he needs to work harder on style. I spent all day yesterday depressed and defensive and I just need to move on and look at this like a kick in the pants.

[timbuck2]

Outstanding post! So, the weighting occurs when the number of Dems, Repubs and Independents is to be included in the sample is determined? Therefore, Newsweek did not take the
345 Republicans, 364 Democrats, 278 Independents and then weigh their responses based on a predetermined formula like yours of 37, 33, 30 for party affiliation?

This seems hard to believe. How can Newsweek say that with the passage of just several weeks, the likely party break down for the U.S. has shifted so dramatically?

Still confused.

-T

[Blake#1]

There is an old explanation of American opinion that states: All issues are divided into 3 parts, 1. those for something, 2. those oposed to something, 3. those that wait to see or hear what is getting the most attention. They want to be with the majority. They wet their finger and hold it up to test the "wind". These are the so-called "independents". Don't knock them, they may be voting for Bush.

[longshadow]

"In a three-way trial heat including Ralph Nader/Peter Camejo, among registered voters ....." [emphasis added]

Don't waste your time analyzing this poll: it isn't worth the paper it's printed on. It a poll of registered, as opposed to likely, voters.

Voters who aren't likely to vote... aren't likely to vote! Hence, their opinion is irrelevant to the likely outcome of the election.