CBS News poll released Monday 8/2: 48-43-3 Kerry, up from 45-43-5

[papau]

http://story.news.yahoo.com/news?tmpl=story&cid=694&ncid=703&e=3&u=/ap/20040802/ap_on_el_pr/presidential_polls

A CBS News poll released Monday found the Democratic ticket slightly ahead of Bush-Cheney, 48 percent to 43 percent with Nader at 3 percent. The three-way race in early July was tied with Kerry-Edwards at 45 percent, Bush-Cheney at 43 percent and Nader-Camejo at 5 percent.


The Newsweek poll of 1,010 registered voters was taken Thursday and Friday and the CBS poll of 881 registered voters was taken Friday Saturday and Sunday. Both had margins of error of plus or minus 3 percentage points. The CNN-USA Today-Gallup poll of 763 registered voters was taken Friday and Saturday and had a margin of error of plus or minus 4 percentage points.

[cspiguy]

1. rah rah.

I have heard whispers that the ABC poll will correctly show the long awaited bounce. Hope they get it right this time with CNN/USATODAY showing a 3 point negative bounce.

[Brotherjohn]

4. Look at this thread. CNN/Gallup actually show Kerry with a 5pt LEAD!

...among "registered adults", and a 3pt LEAD among "national adults".
http://www.democraticunderground.com/discuss/duboard.php?az=show_topic&forum=132&topic_id=592356

Where Bush has the lead is among their subjectively-assigned "likely voter" category.

Otherwise, the CNN/Gallup poll more or less agrees with most others (you won't hear CNN saying that, though).

[Frodo]

6. You got that a tad backwards.

It's a five point lead among "national adults" and three among "registeres voters"

You can ignore (and they should not even report) "national adults". It includes people who can't vote. Their opinion is not particularly relevant unless "likely" questions indicate they are likely to go vote.

[Brotherjohn]

7. My bad. "National Adults"=5pt lead. "Registered Voters"=3pt lead. But..

... even that 3pt lead is a 6pt swing from Bush's 3pt lead among "likely voters".

I would argue that "registered voters" is a more reliable predictor of what the results may end up being than Gallup's "likely voters", unless they have some hell of a crystal ball as to who will be voting this year. It certainly is more objective.

I would also argue that "National adults" is not meaningless at all. Remember, these are just the 1000 or so people they are actually polling, in an attempt to extrapolate the mindset of the entire nation. If the election were going to be just among these 1000 people, then yes, you could throw the "non-registered adults" right out the window.

But it is not. "National adults" is a good indicator of the overall preferences of, well, national adults. You assume the preferences among this small group can not only predict the results of the election (which is the idea, after all), but predict the breakdown of preferences between registered and non-registered adults. That is starting to get a bit far afield from what the poll is designed to find out.

Unless Gallup can demonstrate that "national adults" is somehow a less effective predictor of the eventual results than their "likely voter" category, then I would argue that it has at least as much meaning (if not more, since, like "registered voters", it is also more objective than "likely").

[Frodo]

8. Don't sweat it. But don't fall for the "underdog syndrom" either/

As soon as one side starts to get desperate, you start to see excuses for why a certain poll doesn't mean what it says. And one of the first arguments is this bit about "likely voters"

There's no real magic. It varies from firm to firm, but it's basically "Mrs Jones, do you intend to vote this year?" combined with questions (sometimes) about whether you voted in previous elections.

People who aren't even REGISTERED to vote are worthless polling samples. They don't CARE.

The "registered/likely" gap is more likely a sign of potential for Kerry. He needs to give those people a reason to come out to vote.

[Brotherjohn]

9. Not really underdog syndrome, and not desperate at all. Actually...

... I'm very confident. I look at all the polls in sum and I believe, in reality, Kerry is ahead by about 5%. If the election were held today, i think he'd win handily. I see polls like the CNN/Gallup one that are outliers, however, and it's more a mental exercise kicking in saying "what's different about this one?".

I also believe Zogby underestimates Bush approval (at least relative to other polls), for instance, by the way it asks the question and groups the categories to determine "approval" and "disapproval".

I agree with your point about potential for Kerry. However, I think you still underestimate the significance of "national adults" and overestimate that of "likely voters". If a different 1000 people were polled, for instance, perhaps more Kerry people would fall into the "likely" and "registered" groups. There is nothing short of a pretty small sample size here (especially considering non-registered adults) to show that "likely" voters have such a strong preference for Kerry.

[WI_DEM]

2. a three point bounce

but the bounce was supposed to be small per many pollsters due to the closeness of the election.

[diplomats]

3. And a 5-point lead is good

This just proves that CNN poll was bunk. Every other poll shows Kerry-Edwards with a 5-8-point lead.

[Frodo]

5. Not particularly

It puts the :bounce: range at -3% to +4%. Not particularly encouraging since it represented our "lower expectations" spin prior to the convention.

But it isn't too dangerous at this point. Kerry still leads, and Bush is unlikely to get much of a bounce.

[papau]

10. range at -3% to +4% ?- He gain 3 points, Bush result is not independent of

Kerry result.

Up 5 with MOE of 3% amoung Registered is a spread of up 2 to up 8 for a 95% likelyhood result.