More proof that the CNN poll is garbage

[lancdem]

I just checked http://www.pollingreport.com and saw that the CNN poll of likely voters has the Dems and Repukes TIED at 47 in the generic ballot.

Every other recent generic ballot poll shows the Dems with a lead - and often a sizable one. This just shows they oversampled Repukes among their likely voter survey, as if we didn't already suspect that.

[Inland]

1. www.pollingreport.com

is where I do my poll of polls. CNN/Gallup is looking like an outlier--so to the extent you believe polls, zogby has kerry up by six and Newsweek by seven and ABC by two.

[jjmalonejr]

2. The story the media is writing, though, is NO BOUNCE

That may be true (Kerry did not receive the bounce that candidates usually do after the convention), but the key is to analyze WHY there was no bounce when history tells us that it usually occurs.

The electorate is polarized, with very few undecideds, so there was essentially no bounce to be had. That's why Matthew Dowd pushed this idea of a 15-point bounce...so they could crow about it when it didn't happen.

I imagine that internally, though, they are quite worried about the erosion of confidence in GWB, his low re-elect numbers, his sinking approval numbers, and John Kerry's rising numbers.

That, coupled with the unreported story of this election: the record turnout of new or historically disenchanted voters, should be a MAJOR cause of concern for BushCo. Most of these people will be coming out to vote for Kerry.

[lancdem]

6. Plus, Kerry was way ahead of where challengers

usually are entering their conventions.

[trogdor]

12. The media wants it kept close.

It makes for better copy. Everything's about drama, and if Kerry pulls away by ten points, there ain't no mo' drama.

[Frodo]

13. "No Bounce" is ok. It's the reporting of a "negative bounce" that

is innacurate and unfair.

[TexasSissy]

3. The CNN poll was updated today to incl. 8/1.

[lancdem]

5. I know

The poll number I talk about is from July 30-Aug. 1. I'm saying it's totally inconsistent with other polls measuring generic ballot preference. It's certainly an outlier.

[oasis]

4. CNN can bounce this:

:puke:

[Kokonoe]

7. past bounce

http://abcnews.go.com/sections/politics/Vote2004/dnc_convention_bounce_040725.html Convention Bounces
Net Change
2000 Rep. Bounce Bush +6 Gore -5 Bush +11
2000 Dem. Bounce Gore +10 Bush -9 Gore +19

1996 Rep. Bounce Dole +8 Clinton -7 Dole +15
1996 Dem. Bounce Clinton +4 Dole -1 Clinton +5

1992 Dem. Bounce Clinton +14 Bush -16 Clinton +30
1992 Rep. Bounce Bush +6 Clinton -10 Bush +16

1988 Dem. Bounce Dukakis +7 Bush -4 Dukakis +11
1988 Rep. Bounce Bush +6 Dukakis -5 Bush +11

1984 Dem. Bounce Mondale +9 Reagan -7 Mondale +16
1984 Rep. Bounce Reagan +4 Mondale -4 Reagan +8

1980 Rep. Bounce Reagan +8 Carter -5 Reagan +13
1980 Dem. Bounce Carter +10 Reagan -7 Carter +17

1976 Dem. Bounce Carter +9 Ford -7 Carter +16
1976 Rep. Bounce Ford +4 Carter -3 Ford +7

1972 Dem. Bounce McGovern 0 Nixon +3 McGovern -3
1972 Rep. Bounce Nixon +7 McGovern -1 Nixon +8

1968 Rep. Bounce Nixon +5 Humphrey -9 Nixon +14
1968 Dem. Bounce Humphrey +2 Nixon -2 Humphrey +4


For consistency, these results are all among registered voters. 1992-2000 polls are from ABC News; 1968-88 polls are by Gallup. Earlier polls weren't done frequently enough to track the convention bounce reliably.

[Beetwasher]

9. Past Bounces Are MEANINGLESS and IRRELEVANT

None of the challengers in the past were AHEAD of the incumbent going into the convention. There was essentially no room for a significant bounce to be had by Kerry.

[Frodo]

14. That's not true

I mean, yes, previous bounces mean very little. All that matters (and it matters only a little bit this year) is the RELATIVE bounce for each candidate. Bush is unlikely to get much either.

But you were incorrect on your assumption. More than one challenger was ahead going in to the convention period. You also have to realize that there's a "bounce" attributed to selecting a running mate, and that historically happens closer to the convention (or AT the convention) so a part of Kerry's lead going in to the convention WAS "bounce" from Edwards.

As an example, Carter was up on Ford by almost 20 points BEFORE the convention. It was around 2-1 after the Democratic convention and around a 10% lead after both conventions were finished.

Of course, Carter went on to win it. :-)

Dukakis was ahead of pappy Bush going in to his convention (I think by 6%) and led by 17% coming out. Mondale led Reagan earlier in the year (but not RIGHT before the convention) and come out with a lead.

Of course they lost.

Clinton was in THIRD by some estimates and came back to win.


The moral of the story is that when we say "polling this far out doesn't mean much", it's because it DOESN'T MEAN MUCH, not because were whistling past the graveyard

[lancdem]

16. You're right about the VP bounce

Kerry's pollster, Mark Mellman, said before the convention not to expect a bounce because he'd already gotten much of it from his choice of Edwards. I just want to make a small correction to your post: Dukakis was not a challenger in 1988, though Carter obviously was in 1976.

[Beetwasher]

17. I Was A Bit Too Absolute

in my comment about the challengers...You are correct...However, my point remains about the insignificance of Kerry's supposed "non-bounce", there was really hardly a bounce to be had.

In the Dukakis/Bush election however, there was technically no incumbent, so it's a bit of a different scenario.

The Carter/Ford race is a better example (even better because Ford too was not actually elected President). Carter was ahead going into that convention against an incumbent, and as you pointed out, Carter won...That bodes well for Kerry...

[MallRat]

8. Why aren't competitors laughing at CNN's ineptitude?

Good catch on the generic ballot numbers... even Republican pollsters like Rasmussen have the Dems ahead by 4-7 points in the generic Congressional ballot question.

If I were CNN, I would have mentioned the poll and then pretended it never happened. That's what most networks do when their poll comes in out-of-whack with everyone else's. They just look goofy right now.

-MR

[DemocratSinceBirth]

10. They Have Declared War On Us

And the flawed Gallup Poll which is intended to shore up Republican support is their Weapon of Mass Destruction...

[DemocratSinceBirth]

11. Kick

[jmknapp]

15. Poll noise != poll news

I find the poll conspiracy theories shallow. The Gallup poll is fine. It's just that there is a lot of statistal noise in these polls. With a 3% margin of error there's a one-in-ten chance that a tied poll will result even if the real margin is 6% (take 3% from one and give it to the other).

[Beetwasher]

18. If That Were True Then Why Don't The Errors

Ever happen in Kerry's favor?

The timing of this supposed "outlier" and "anomalous" poll is very suspicious. Right after the convention so it can be used to reflect negatively on Kerry and his supposed "non-bounce". Sorry, I'm not buying your coincidence theory, especially considering Gallup's history of releasing questionable polls, with anomalous results at opportune times that always seem to help Bush's campaign.

[DemocratSinceBirth]

19. I'm With You...

CNN used that poll like a hammer to nail the Kerry campaign for not getting a bounce from his convention while ignoring contracicting polls....

CNN is Fox in drag.......