MyDD :: GMAB Newsweek

GMAB Newsweek

by Jerome Armstrong, Sat Oct 30, 2004 at 04:10:49 PM EST

The just-released Newsweek poll, Breaking for Bush?, contains the usual pro-Republican makeup (poll details) that we've come to expect in the LV model done by certain polling operations that don't weight by Party ID. It's amazing what sort of story the media can create with just 1117 individuals of their choosing. In their just released poll, which the conservative mouthpieces are using to tout a "late break to Bush" as a result of Osama, and a 6% lead over Kerry among "likely voters" here's the facts.

Among 1005 registered voters, Newsweek has Bush leading Kerry 48-44.

From the RV's, Newsweek has culled 882 likely voters, where Bush leads Kerry 51-45.

Here's the voter ID of those 882 LV's (2000 exit poll results):

345 Republicans    34.5%  (35)
316 Democrats      31.6%  (39)
304 Independents   30.4%  (27)

I'm not going to spell out the whole argument that's been made here before by Chris and others here and here. In short, if weighted by Party ID, the Newsweek poll would likely find a dead-heat, or a slim Kerry lead. By under-representing Democrats by 7 percent, Newsweek stacked the deck of LV's to favor Bush; and they over-count Independents by 3 percent. That's where the "late break" by Newsweek is to be found:

Last week Kerry captured independents’ support 52 to 38 percent. This week, for the first time since the debates, Bush has retaken the lead among independents, 47 to 38 percent.

Seen that confirmed anywhere else? Because that's the basis of "the break" to Bush on the cover of Newsweek the day before the election.

Tags: General 2008 (all tags)

Permalink :: 19 Comments :: Post a Comment

right vs. wrong (none / 0)

It never ceases to amaze me that people (you listening W?) keep doing the same thing wrong over and over again, even when they know its wrong. $$$

by shlenny on Sat Oct 30, 2004 at 04:24:04 PM EST

Numbers (none / 0)

I think your numbers are a little off:

345+316+304= 965 Voters and then your percentages only add up to 96.5%

I was trying to figure out the problem but I can't find it.

Regardless, Newsweek is still clearly oversampling Republicans, so its not a big deal

by dbeard115 on Sat Oct 30, 2004 at 04:25:23 PM EST

Re: Numbers (none / 0)

Well, their are MoE's within those numbers too. The underlying misinformation is that Democratic voters are under-represented.

by Jerome Armstrong on Sat Oct 30, 2004 at 04:40:48 PM EST
[ Parent ]

Re: Numbers (none / 0)

The numbers don't add up because of the weighting...from the notes below the poll results:

Data is weighted so that sample demographics match Census Current
Population Survey parameters for gender, age, education, race and region.

Reported sample sizes are unweighted and should not be used to compute
percentages.

One thing you CAN see in the raw numbers is the fact that there are more red state people than blue state people polled:

321 Republican states (plus or minus 6)
397 Swing states (plus or minus 6)
287 Democratic states (plus or minus 7)

I assume they define "Republican" states as states that Bush won in 2000. If so, this is pretty bogus, considering he lost the popular vote.

by KMA on Sat Oct 30, 2004 at 04:47:18 PM EST
[ Parent ]

the poll doesn't add up (none / 0)

46% approval rating (easily lower than many of the tracking polls that have the race closer) but a 6 pt lead? Their numbers are contradictory.

by asf6 on Sat Oct 30, 2004 at 04:29:22 PM EST

Why does the Electoral Vote Map? (none / 0)

Have Illinois as Republican? Or Ohio and Michigan for that matter?

by dmsdbo on Sat Oct 30, 2004 at 04:43:29 PM EST

Re: Why does the Electoral Vote Map? (none / 0)

LOL, I am getting the map updated to have "uncalled" so that MyDD can call the battleground states on election night, and while it's getting worked on, there's a bit of testing...

by Jerome Armstrong on Sat Oct 30, 2004 at 04:58:03 PM EST
[ Parent ]

Re: Why does the Electoral Vote Map? (none / 0)

That's funny! Good idea, and it seems to be good to go from my end. While your at it, should Hawaii be put in the undecided column for the hell of it (at least until a poll with some sense pops up)?

by dmsdbo on Sat Oct 30, 2004 at 05:00:11 PM EST
[ Parent ]

Re: Why does the Electoral Vote Map? (none / 0)

On election day, they will all go to uncalled, and we'll work from there, should be fun!

by Jerome Armstrong on Sat Oct 30, 2004 at 05:03:35 PM EST
[ Parent ]

Another republican over-sampling (none / 0)

If my math is correct, I get Bush 48 Kerry 46 after applying the 39/35/27 model. A dead heat.

by Sin City Guy on Sat Oct 30, 2004 at 04:54:17 PM EST

The media poll distortions continue .... (none / 0)

...... "Seen that confirmed anywhere else? Because that's the basis of "the break" to Bush on the cover of Newsweek the day before the election." ...

What I am constantly seeing and hearing in the corporate media is the constant attempt to create the impression (control the narrative) to imply an exorable Bush win.

These manipulated poll data results (or falsely normalized poll results) are just more of the same.

Many knowledgeable people point out that concentrating on polls and polling data is corrosive of the process of democracy.

Clearly, creating a false impression of Bush's popularity has been a defining technique for Rove - Bush since 1999, and the media continues to cooperate.

It worked for them in 1999 / 2000, no one has the means to counter a symbiotic GOP - Corporate broadcast media collusion like we see consistently at work in the reporting of misleading polling data.

I think we have to do something about this as a society, I think it is that important an issue.

Some democracies ban the publication of poll results within a certain period before an election.

Obviously, we are not ready for that as a society.

But we need to do something.

Perhaps a party based polling peer review board to (quickly) respond to these deliberate distortions would work.

And added to this would have to include the strategy that all media people for the campaign must absolutely refuse to answer or comment on any question whatsoever put to them in terms of a polling result.

For one, Paula Zahn's entire interview approach is the promotion of these false poll data "facts" to absolute certainties and she consistently uses this in attack mode against the Democrats.

She is not at all even handed in the application of this corrupt reporting or interviewing technique.

Paula Zahn is worse in fact that Chris Matthews who at least has publicly stated his political preferences and leaning. Paula Zahn maintains the pretence of objectivity.

The campaign polling review panel would be referred to for any polling question, if that approach was followed with discipline, it would work and it would take one more tool away from liars like Rove and Co and their sycophants in the corporate media.

CNN yesterday used a Jeff Greenfield report on Paula Zahn's program to work on this same theme; an exorable Bush lead in the polls leading to an inevitable Bush win on 11/2, with a little of the "Kerry can't close the deal" perspective I stated they were going to use a few months ago.

This is a serious issue and we need to do something about it.

Just pointing on HOW they do it time after time, what statistical technique they are using at the moment is not a sufficient response.

by leschwartz on Sat Oct 30, 2004 at 04:58:56 PM EST

Backfire (none / 0)

They are going to find their spurious polling backfiring. I for one, felt sort of maxed out on political involvement, phone banking for Kerry. Now, I'm energized again, and volunteering again. This is great for GOTV efforts.

Let them eat their polls. We're going to work on the only poll that matters.

by Alan S on Sat Oct 30, 2004 at 05:01:10 PM EST

Re: Backfire (none / 0)

Alan,

Its worked for them before, it was the most prominant part of the 2000 media campaign.

You can't rely on your personally felt preferences carrying the day.

Unfortuantely people have been, are now, and will be swayed by these techniques, else why would it be so fundamentally a part of the Rove - Bush media stratagy.

We can't just simple-mindedly say GOTV in response to these media distortions.

GOTV will not motivate people if they have already been conditioned day after day to believe in the opposition's inevitable win.

This is just one more voter supression tool being used by the GOP and we are fools if we let them use the GOP biased corporate media, all for free, to do this election after election.

by leschwartz on Sat Oct 30, 2004 at 05:19:19 PM EST
[ Parent ]

Slight improvement in the WaPo tracking poll today (3.00 / 0)

Bush up 49-48 among LVs, after Bush 50-47 yesterday.

Kerry up 48-47 among RVs, after Bush 48-46 yesterday.

I like those RV numbers.

by Haggai on Sat Oct 30, 2004 at 05:07:21 PM EST

Oh, well, Newsweek must be right (none / 0)

Even every other poll and certainly more reputable ones show Kerry with the Big MO(Rasmussen, Zogby, and WaPO).

by jp2 on Sat Oct 30, 2004 at 05:12:59 PM EST

here's the thing (none / 0)

I'm quite up on the party weighting argument by now. But, what are the possibilities that they're closer to right than we think they are? All comparisons right now are to 2000. But in 2002, we were all really surprised on how things broke on election day. I remember a lot of confidence only days before the 2002 election, and then Republicans swarmed to the poll. Is it possible that there really are more Republicans than Democrats, and that the turnout figures in 2000 were because of depressed GOP turnout? In 2002 and this year, the GOP has that 72-hour thing. Couldn't that make a big difference?

by tunesmith on Sat Oct 30, 2004 at 05:17:15 PM EST

Bull Sh*T (none / 0)

I commented on this one earlier and I wrote and called Newsweek , NBC, and MSNBC about this blatantly false reporting. If I polled Bush's base and only got a 4 point lead for him I'd be writing the opposite "Bush In Trouble, Big Trouble". We need to hold these morons accountable. Everyone and I mean everyone knows we got a surprise waiting for them come Tuesday.
As Eminem says..." Mosh Now or DIE"....
Don't forget Eminem is on SNL tonight performing"Mosh"...

by joedemocrat on Sat Oct 30, 2004 at 05:28:06 PM EST

WTF...? (none / 0)

"Last week Kerry captured independents' support 52 to 38 percent. This week, for the first time since the debates, Bush has retaken the lead among independents, 47 to 38 percent."