Last CNN Poll: A Dem Tidal Wave

By Big Tent Democrat, Section Elections 2006
Posted on Mon Nov 06, 2006 at 07:45:33 AM EST
Tags: (all tags)

Share This: Digg!

The most recent of the generic polls, the one furthest from John Kerry, predicts a huge Dem tidal wave:

The percentage of likely voters who plan to vote for Democrats in Tuesday's congressional elections increased in the past week, and those voters supporting Democrats also seem less likely to change their minds before casting ballots, according to a CNN poll conducted over the weekend. Democrats hold a 58 percent to 38 percent advantage over Republicans among likely voters in the survey released Monday morning, compared to the 53 percent to 42 percent advantage reflected in the poll a week ago.

And Bush's approval ratings suffered due to his high visibility:

President Bush's popularity took another dip over the past week, and four out of 10 likely voters said their disapproval of the job he is doing will impact how they cast their congressional ballots on Tuesday, according to a new CNN poll. Bush's approval rating fell to 35 percent, with 61 percent of those polled saying they disapproved of the way he is handling his job as president, in the survey conducted on Friday, Saturday and Sunday. This represents a two point decline in Bush's approval rating compared to the CNN poll conducted a week earlier and it is four points lower than the survey taken two weeks ago.

Sending Bush out may have energized the GOP base, but it seems to have deepened the anger of Dems and Independents as well. A two-edged sword for the Republicans.

< Is It Tightening? | In Praise of Barney Frank >

Legal

Nothing on this site should be construed as legal advice. TalkLeft does not give legal advice.

TalkLeft is not responsible for and often disagrees with material posted in the comments section. Read at your own risk.

Last CNN Poll: A Dem Tidal Wave | 11 comments (11 topical)

Readers Must Login or Create an Account to Comment

A two-edged sword for the Republicans. (none / 0) (#1)
by Edger on Mon Nov 06, 2006 at 08:23:01 AM EST

From Charlie Cook, November 3/06:

In the House, it would take a miracle for the GOP to hold onto their majority. The losses look very likely to exceed 20 seats, and a 20- to 35-seat loss is most likely, but we would not be surprised for it to exceed 35 seats. The vulnerable GOP seats are there, the wave is there, maybe it happens, maybe it doesn't.
...
Many have commented, quite correctly, that the biggest variables are turnout levels among independents and Republicans.
...
It is impossible to determine before an election what the turnout levels will be among the various groups. Polling has suggested that Democratic voters are extremely motivated while Republicans are more disillusioned than they were in 2002 and 2004, and anecdotally there certainly are reasons for Republican voters to be despondent, no matter what faction of the party they belong
to...
...
The bottom line is that at this stage, Republicans should consider themselves lucky if their net losses stay in the 20-25 range in the House, four or five seats in the Senate, and between five and eight governorships. It would be a tough election, losing their majorities in the House and governorships, but it would fall short of the devastating losses that are possible. But the chances of this thing going bigger -- far bigger -- still exist

We'll have some real numbers to measure Cook's analysis against tomorrow night.

While numbers are only part of it, let's hope they represent a tidal wave of mindset change as well...

A Sword of Damocles?

I'd feel a lot happier about numbers like this.... (none / 0) (#2)
by Avedon on Mon Nov 06, 2006 at 11:24:27 AM EST

We could see them by district - and color. It doesn't matter if more people are saying they will vote for Dems if they're doing it in districts that are already blue. What matters is how people vote in places that have been going red in the past. We could treble the actual number of people voting Dem without gaining a single seat.

: What matters (none / 0) (#3)
by Edger on Mon Nov 06, 2006 at 11:31:02 AM EST

What matters is how people vote in places that have been going red in the past.
Also from the Cook report:

Perhaps the most remarkable thing about this election is that Republicans are having to fight and spend money in states and districts where few Democrats have dared tread in recent years

Be cautious. (none / 0) (#4)
by Gabriel Malor on Mon Nov 06, 2006 at 11:48:44 AM EST

There are a few reasons to question the latest CNN poll.

First, and most importantly, this was a poll of "adult Americans" not likely voters. Also, it was conducted over the weekend--traditionally, the least representative polling period.

At this stage of the game, this is little more than a "buck-up Democrats" poll following on the heels of the more accurate and less positive for Democrats polls conducted earlier last week.

That is my second point. This poll differs significantly from others conducted last week.

First came the WaPo/ABC poll conducted Wednesday through Saturday (though also of "adult Americans" it actually separated its results into those who self-identified as "registered voters"). It found:

Generic Likely Voter, Dem +6 (down from Dem +14)
Right Track/Wrong Track, 39/59, up from 32/66
Who do you trust on Iraq, Dem 42, Rep 42, up from Dem 48, Rep 40

Then came the Pew poll, also conducted during the week, which revealed among likely voters

Generic Ballot Likely Voter, Dem +4 (down from +11)
Leaning, or Flat Undecided Voter, 19%

And then came Gallup, which also polled likely voters. It found

Generic Likely Voter, Dem + 7 (Down from +13 two weeks ago and +23 a month ago)

Finally, CNN has not released the internals. Meaning we have no idea how many Republicans, Democrats, or independent voters were polled. Nor do we know which of them were registered or likely voters.

Democrats have reason to be optimistic--after all, the polls still point to good things for them. On the other hand, the polls pointed to Democratic wins in 2002 and 2004 and we know how that turned out.

There are many reasons to question the accuracy of polling, but chief among them is the fact that they have failed--for the past six years--to actually track votes. Polls are often distorted by the fact that Republicans are both less likely to respond to pollsters and more likely to mess around with them when they do bother to participate.

That was the lesson of the disasterous exit polls in 2004--a lesson which has likely gone unlearned. It's so much more fun to declare that lost elections are simply unfair and point to pre-election and exit polls as proof of "funny business." Keep in mind that all polls except for the one that matters on Election Day are only 95% CI--that's a plus or minus 3%, generally.

And Fox? (none / 0) (#6)
by Big Tent Democrat on Mon Nov 06, 2006 at 01:58:59 PM EST

Re: the Fox Poll (none / 0) (#7)
by Gabriel Malor on Mon Nov 06, 2006 at 02:42:57 PM EST

The weekend Fox poll results can be found here (PDF).

This one looks pretty good for Democrats, though it also is a weekend poll. It's 900 "likely voters" so it's worth taking note.

The key findings are that those polled would vote for the Democratic candidate in their district 49% to 36%. That's Dem +13, (better than the previous +11). The key is the 15% who declared that they were undecided. As ever, the margin is 3%, so this is well outside of any statistical difficulties.

The other big discovery (at least for me) is that 49% of the self-described Democrats polled claim to have already voted compared to 34% of Republicans. I would be interested to know if that's the case generally, and I would speculate that voter turnout tomorrow (rather than absentee or early voters) will be the deciding factor in who walks away with the win. Based on this poll, it's the Democrats race to lose.

Again, I want to extend my earlier remarks regarding polls in general. And I would say the same were the results reversed and Republicans were showing ahead: be cautious.

Phone polling is becoming increasingly difficult. These days the results include only those who have listed numbers and are willing to answer 30 minutes of questions. That's a pretty select group. It largely leaves out the wealthy and the upper-middle class (and obviously students, but they don't vote in numbers, anyway). It also tends to select away from those who are busy or unenthusiastic about this election.

For those reasons it oversamples a rather peculiar group: those who did not care to opt onto the do-not-call lists, who also have the time and inclination to talk to pollsters.

Finally, this poll reports that the results are based on 900 likely voters. It does not tell us how many calls were made before the 900 LVs were reached (I'm assuming those calls were thrown out of the final results--it would be miraculous and unbelievable for Fox to cold-call 900 people who all said they were likely voters). Overall, turnout is probably going to be around 60%. I would be interested to know just what percentage of those Fox called claimed to be going to the polls tomorrow. Sadly, Fox has not released this data.


polling (none / 0) (#8)
by eric on Mon Nov 06, 2006 at 04:14:39 PM EST

That's a pretty select group. It largely leaves out the wealthy and the upper-middle class

How? It seems to me that the wealthy and the upper-middle would be exactly the type of people to have a land-line phone and answer questions.


Re; polling (none / 0) (#9)
by Gabriel Malor on Mon Nov 06, 2006 at 04:24:23 PM EST

Those groups are most likely to be on the do-not-call lists. They are also the most likely to have no time for pollsters--especially on weekends.

Those most likely to have time for pollsters are the elderly and the unemployed.

Those most likely to have the inclination to talk to pollsters are (this season, at least) Democrats.


Do Not Call (none / 0) (#10)
by Big Tent Democrat on Mon Nov 06, 2006 at 05:33:17 PM EST

does not apply to polling.


: ...I suspect you're right. (none / 0) (#11)
by Gabriel Malor on Mon Nov 06, 2006 at 05:42:58 PM EST

Damn...I'd forgotten that. The only restriction that is in effect then is unlisted numbers (not counting those who've given out their number with contributions and such).
Still, it amounts to about the same. The rich and the upper-middle class are much more likely to have unlisted numbers than other income groups.

Actually, much like when bush was trying... (none / 0) (#5)
by Bill Arnett on Mon Nov 06, 2006 at 12:29:59 PM EST

...to get rid of Social Security, EVEN PLAYING TO PEOPLE WHO HAD TO SIGN LOYALTY OATHS TO HEAR HIM SPEAK, the more he spoke the fewer people believed him, and the more his poll ratings dropped. The same thing happened here.

About three weeks ago he had 36% support on Iraq. He support for handling Iraq now stands at 29%.

I just love it when people are so stupid that they don't even realize what a joke they have become, even to their supporters!

Nice job, mr. boosh. Keep at it, you can still drop 29% more!