« Kerry Holds Narrow Lead in Economist Poll | Main | About That Vanishing Gender Gap »

Dead Heat in New Democracy Corps Poll

John Kerry and George Bush are tied at 49 percent of nation-wide LV's, with 1 percent other in a poll conducted Sept. 19-21 by Greenberg Quinlan Rosner Research for Democracy Corps.

Comments

Great news, obviously. Hey, could you-all put "warning: PDF" or something when you link to PDFs? Some polls are PDFs (like the DCorps one), some are just websites, and some of us have exactly enough time at work to read websites with numbers and Ruy's great analysis, but not to wait for Acrobat to load. Just askin'.

Idea for future posts: (1) how are we polling in battleground states relative to how Gore polled there in late September 2000? We've fallen into the habit of comparing state polls to 2000 election results, which favored Gore relative to the latest state polls. Apples-to-apples 2000-2004 comparisons might fuel optimism. (Though with the new batch of national polls, we should have optimism to spare.) (2) Kerry is underperforming Gore with African-American voters, it seems-- but by how much? Is this a real problem, or an illusion? (For example, if African-American voters are always late deciders anyway, it shouldn't matter that we've got more undecideds in Sept polls than Gore did in the actual election; if more African-Americans are actually voting for Bush this time, that has to be a problem-- get well soon, Big Dog!)

It's worth noting that Democracy Corps also asked those sampled about their 2000 vote. The results were Bush 51, Gore 43. OK, Ruy, how can this be? Could a poll showing a dead heat still be significantly oversampling 2000 Bush voters? Could the electorate somehow change drastically in four years, with many more Gore voters than Bush voters dropping out? Or are there a lot of people who are incorrectly reporting their 2000 votes? All of these answers seem implausible, but at least one of them has to be right.

poll is very encouraging. I wonder what your (EDM) current projection is for electoral college result? I am hopeful for Minnesota, but scared about the possibility of losing this state to the righties.

Also wondering about the track record of these various pollsters. Do they have any bias (political or procedural) that throws off their accuracy?

true believer

This is terrific news, and it was echoed on nbc this morning as an opener. Great! My take on the numbers is that Kerry didn't lose support as much as Bush gained superficial support after the NY convention. Since then, Bush's increasingly tone deaf spouting off about how great things are in Iraq, "marching towards democracy" has hurt him badly. The latest wind-surfing ad also hits the wrong note - light hearted and funny, but they're talking about a war where our troops are dying! Bad. I also increasingly hear from the pundits how the Kerry flip-flop answer from Republicans to legitimate questions about Iraq is wearing very thin.

Something's changing in this race - I can feel it.

I raised the problem of oversampling of 2000 Bush voters a while ago on a different thread, and since then it's been reported in connection with several other polls.
Someone answered me then that it may be partly a bandwagon effect: people are mistakenly remembering or claiming to have voted with the winner. I think that may be part of it, but then, I should think at least some people who did vote for Bush would be too embarrassed to admit it!
Why doesn't someone do a poll and weight it by reported 2000 vote, rather than by party identification? If there's still a significant number of Dem. identifiers, in the south especially, who haven't voted for a national Dem. ticket since the days of Jimmy Carter, weighting by 2000 vote would produce a better prediction than doing it by party.
At least I'd like to see someone try correlating all these things: claimed 2000 vote, party identification, 2004 preference.
BTW, the mis-remembering problem reminds me of a study of church attendance in a certain area, in Ohio I think, which found that the number of people claiming to have attended church the previous Sunday was nearly twice the total turnout reported by all the churches in the area. Someone should do the same with voter turnout as well as preference.

Once again, look @ Q92 - voted for who in 2000

Bush 51, Gore 43, Nader 3

Diff @ -8 is one point better than -9 last time.

Is the pool spiked or is this "of course I voted for the winner"?

People tend to claim, or believe, that they voted for the winner. Polls that show more Bush 2000 voters are unlikely to be suffering from sampling bias for that reason.

I wondered if anyone had any thoughts about the television rollback by Kerry. I understand why their doing, but it strikes me as suspiciously like Gore's decision to pull out of states where it turned out he came quite close.

Regarding some recent Wisconsin polls showing a big lead for Bush, please see the following piece I wrote up as a Kos diary, if you're interested:

http://www.dailykos.com/story/2004/9/23/1369/38422

Regarding the recent Wisconsin polls that show Bush up big, please read the following piece I wrote up as a Kos diary, if you're interested:

http://www.dailykos.com/story/2004/9/23/1369/38422

Wondering about the hispanic/latino percentage, which seems below their historic participation (or is it just me?)

I think it is quite common that these kind of questions oversample the incumbent. There seem to be a significant number of people reluctant to be associated with the former "loser".

But this is also good news. Because that shows this poll doesn't oversample Democrats, does it?

This is great! Kerry is turning the tables. Come on, new CW!

Two quick comments:

1) People don't always remember who they voted for. Most people like to say they voted for the winner, thus the skew in favor of Bush.

2) I have overseas friends, and each one of them hates Bush. I use "Hate" for a reason. We in the US forget how PC our dialogue really is. Most of the rest of the world despise the President. My guess is that vote is going to be much bigger this year, and it will tilt Democrat.

I'd like to know why DC gets a significantly higher number of Nader backers? All the other polls show him at 1% while DC has him at 5%.

Nader won't get 1% on election day...

The talk is already abuilding out there of Kerry's BIG SURGE in the polls...

Sheesh...the race hasn't really moved all that much as DR readers know full well and yet..Kerry supporters in the blogosphere are energized, they're energized at SF Demo Hdq, and the pudits are just waiting for the next Gallup poll to pronounce..

I read somewhere that in the NES time series, voters follow polls more than any single issue in a presidential campaign

Ain't it the truth

Funniest thing is, the Wall Street Journal, not exactly a bastion of progressive liberal ideas, published a poll today showing the race tied, and even had a lead story about Kerry catching up.

> My guess is that [oversees] vote is going to be
> much bigger this year, and it will tilt Democrat.

If so, watch for the right-wing hate Wulitzer to pick that out as evidence of "cheating"

> I'd like to know why DC gets a significantly
> higher number of Nader backers?

It's a good thing. People there know that Kerry will take DC, and so are voting their true preference, rather than simply making sure to kick out Bush. If Kerry was destroying Bush in the polls, the Nader vote would probably be much higher, because Naderites wouldn't feel pressured to vote for Kerry.

As for the overseas vote, the International Herald Tribune reported Monday that the Pentagon, on the eve of absentee ballot deadlines, has begun denying ex pats access to the Federal Voting Assistance Program website, citing vague and unspecified "security" concerns. Soldiers can still use the website, but the ISPs of the heavily Democratic leaning ex pats are being denied access. This after a year of ex pats being encouraged to rely on the website for getting their ballots. There are about 6 million Americans living abroad and this throws a monkey wrench into their voting plans. We need to take into account that these sorts of dirty tricks are going to be used all over Red America to supress traditionally Democratic voters, such as blacks in Florida. Kerry is going to have to get substantially higher in the polls to offset this type of anti-democratic attrition.

ABC has President Bush up by 10 points in Wisconsin. A Badger Poll (conducted by the University of Wisconsin) has the president up by 11. This is consistent with strong results in state after state. Can all of these be wrong? Or, could the tilt in republican I.D. be the results of independents now supporting the president, and now calling themselves republicans? Or, could it be the result of a strong campaign ground game?
From ABC News The Note:
http://abcnews.go.com/sections/politics/TheNote/TheNote.html

Registered voters are 6 points more likely to have been contacted by BC04 than by KE04, the poll shows 25 to 19 percent. Sixty percent of those who have gotten Bush calls are supporters, whereas Kerry enjoys the support of fewer than half of those his campaign has reached.

Speaking of Nader...

Just had a discussion about 3rd party candidates with a friend of mine. My argument was that we should have runoff elections to ensure that the candidate who is closest to the views of most voters gets elected. He agreed, but also made the argument that if people voted the way they are supposed to, based on how our system is currently set up, then it would make no difference.

This is a seemingly obvious statement, but is actually somewhat profound if you think about it.

What is accomplished by voting for Nader? If we take people's political beliefs as a straight left to right continuum, people should elect the candidate that falls closest to them on this continuum. Nader voters need to suck up the fact that our political system is a joke and actually vote for an electable candidate. So, my advice to not just Nader voters, but anyone who is voting, is this: Determine where you are on the political spectrum and choose the electable candidate who is closest to you on it. Say you have candidates A, B, and C, candidates from left to right (say in roughly equal proportion). A and B fall slightly more on the liberal side, while C is way off to the right. 45% of people are closest to C on the spectrum, 40% closest to B, and 15% closest to A. Mean or median Averaging brings the "center" of the country somewhere between A and B, but closer to B. Who should win? Isn't this obvious? Doesn't the mode seem like a stupid way of electing someone? I can't wait until we do actually have an 8 way race or something and someone way off to the left or right wins with only about 20% because a small bloc of people are exactly in line with him, despite one of the 8 candidates actually being closer to the rest of the 80% of the country.

Until we fix our system, the sad reality is that in order for the system to actually work the way it's intended is that you have to vote for the electable candidate. Unless you really believe that Bush and Kerry are the same thing (which is hard to believe and even Nader doesn't admit to), the only message that voting for Nader, or any other candidate, sends in a close election is that you don't care if the minority of the country's views are represented by our leader.

I was pleased to hear Chris Matthews last night on Hardball say that someone, "very smart, very smart" was saying to "watch for Kerry to win this by ten points".

One quick comment on who people say they voted for. They DON'T necessarily oversample for saying they voted for the "winner" but rather for whichever way the winds of power and respectability are blowing at the time they are asked. Hence, in the mid 70s, MANY fewer people said they voted for Nixon in 72, landslide winner, than actually did. The question is not whether they won the election, but whether the person is fearful of being identified with them when the question is asked. I suppose the same logic applies to "the fear of God [and respectability]"

The latest Fox poll is out. Bush up by 4% in the three way, 2% in the head to head among LV's.

50% Bush job approval.

Kerry's improving in strength of support. 76% of Bush supporters, 72% of Kerry supporters give strong support.

Friend of mine in his 70s, an amateur historian and steeped in American political, military and social history said he thinks Bush's hard to believe approval ratings are a "courtesy thing". They seem so out of sync with 53% "does not deserve reelection" ratings etc. ANd he's not a yellow dog Democrat. He voted for Reagan and Bush pere, but cannot abide Bush Jr. My comment here was triggered by other postings here that voice skepticism about all the poll respondents who say they voted for Bush in 2000. I think many such persons don't want to appear disrespectful to the monarch. Dunno, mebbe I'm overanalyzing, but the disconnect baffles me.

Yes, I too, heard that cryptic remark by Chris Matthews -- that some very smart person told him Kerry would win. With all his GOP scripted commentary and his dad having been a lifelong GOPer, I sometimes suspect he is a closet Democrat. He was, after all, mentored by the late Tip O'Neill.


Sep 23
Rasmussen has Kerry with a slight lead over Bush in FL 3 days running. Within MoE, but a trend maybe?
He's had him tied in PA now, too.

Stay tuned, lil' wranglers and keep them postcards comin'. An' be sure to eat your spinach, too.

Re 1972 voters' amnesia

In 1974, best bumper-sticker:

Don't blame me, I'm from Massachusetts.

(MA was only state non-Nixon in 1972, which was, lest we forget, the greatest electoral vote victory by any modern president.)

Jeff, as a friend of mine pointed out, the person most closely aligned with my political beliefs is me. And I'm not going to vote for me because I have very little chance of winning.

I just watched a segment on Lou Dobbs about job losses in OH. Hard to believe OH will go for Bush. Maybe I'm whistling by the graveyard. Guess we'll know in 40 days or so.

1) Re: African-Americans. Actually- an article out of a Phildelphia newspaper, and my own anecdotal knowledge says you are wrong about African-Americans. They, and I including me in that, are "pissed" over 2000. What was under report, but which has been reported a lot in the black press is the number of black people who were disenfranchised in Fl. We maybe as a group more socially conservative, but as this issue- disenfranchisement, the draft issue for younger voters and pop cultures involvement- not to mention the more traditional bread and butter issue will produce one fo the largest African-American turn out in a long time. One of the things that angered them about Gore in 2000 was that he failed to argue that we were disenfranchised- behind the scenes a lot of people have been puttin gin a machine to make sure this doesn't happen twice. They are trying to use the gay issue to convince AA's to join them- but that's a mistaken tactic b/c most people say "we can't trust them after 2000."

2) I agree with the post above that says that Bush is making a strategic mistake but not directly answering Kerry on his offense that Bush has created a mess in Iraq. The fact is the reason why the attack by Kerry works I think is that Americans at their gut level don't like when people don't take responsibility for their own actions. And, Bush increasingly is feeling like a guy who is not taking responsibility. In fact, as I argued else where, I believe this should be the narrative that Kerry uses until the election day for Bush- that he's a man who ultimately the buck doesn't stop with him- it stops with everyone else.

3) B/c of public pressure the pentagon has alledgedly openned up their site once more to ex pats. What this means I don't know, but I read it doing my daily google search of news this morning - or maybe it was the yahoo search.

4) Re: Ohio. THe problem with the way many people vote and this explains why some Americans vote the way they do is explained by Molly forget her last name? on that new left leaning radio network. She put it this way for Republican wins- it kind of a kind of hide the ball. Ohio goes to vote against gays getting married, they are voting for huge deficits. they think they are voting for security, they end up voting for a regressive tax structure. in other words, they think they are going in to vote for agenda x, but instead are getting x and y. I have a chrstian friend who is like this. He told me he is thinking of voting Bush b/c he is more in line with his belief system- but I pointed out to him that under our system of governance, his belief system would never become law any My friend of course knows this b/c we both are lawyers. He tells me, but there is more of a tendency for him to vote along Christian lines. He is in other words voting for x although bush is really giving him y b/c when I point out all of the position for which he disagrees with Bush on he is hard pressed to explain his voting decisions other than that Bush is Christian.

Bill Makuch,

Your 4:06pm post makes an excellent Point: How come reputable recent state polls (excluding the heavily Dem ARG & Zogby) show GWB ahead, on average, in WIS by 11, FLA by 7, OH by 10, IA by 7, and basically tied in PA, MN, NJ, NM, ME & OR - Blue states all except FLA & OH??

That is because Democracy Corp is run by DNC affiliated people. Not exactly unbiased. It's an age old trick: They know all other polls show GWB up nationally by small margins up to ten points, and they see some of these trends in their poll - So they jiggle the numbers to show Kerry is tied when he is in fact behind.

I believe we should start focusing primarily on state polls in part because too many national pollsters have agendas. National polls are a barometer of the country's overall mood, but too many dynamics specific to each state (ie whether Nader is on the ballot, job figures, etc.) come into play at this late stage.

And on a state by state basis, excluding ARG & Zogby subscriber internet poll, it's not looking good for Kerry at this point. GWB is ahead comfortably in Red states, while Kerry is struggling to hold his turf in places like PA, WIS, IA, NM, MN & NJ. And Kerry's churlish performance today, berating the courageous leader of a country we liberated, isn't going to help his sagging poll numbers IMO.

Someone addressed the apparent incongruity of over 50% of Americans approving of Bush, while over 50% think the country needs a change.

The real number is the number who think we need a change. As someone pointed out, a few percentile of the public will say they "approve" while they obviously do not.

Bush is going to lose because the number of voters will be a record, and the new voters aren't lining up to vote for Bush.

http://www.foxnews.com/story/0,2933,133281,00.html

Bush is creating fear again as soon as he realises that Kerry is about to surge ahead in the polls.

bruh rabbit raises the same issue as the author of "What's the matter with Kansas?"
There are communities in which the entire community is on Social Security and Medicare, don't have pot to wee in, but vote GOP approx 60%. Clearly when such folks vote again and again against their interests they must be willfully uninformed.
Orwell would have understood.

Some good news from Gallup for a change:

"Bush, Kerry in Statistical Tie in Florida, Gallup Poll Says"

http://quote.bloomberg.com/apps/news?pid=10000087&sid=a5fyeKbmF4ZY&refer=top_world_news

Odd...I didn't hear Kerry refer to the leaders of Japan or Germany today.

Charlie-

You missed my point completely. I said that you should vote for the most ELECTABLE candidate that is closest to your political beliefs. If you were running, and electable, yes, you should vote for yourself.

A run-off system would produce essentially the same effect as if everyone voted for the electable candidate closest to their views.

As it is, unless you're George Bush or John Kerry, you're not electable. Sorry. You should vote for one of them (and preferably John Kerry, but like I said... whoever's closest to your beliefs from the two ;) ).

Smooth Jazz-

I'm not sure where you get Zogby and ARG being "Democratic". Are you just pulling this out of nowhere, or does it actually say somewhere that they're party affiliated?

The WSJ today basically said that in general, polls with more republicans in the sample were polling with Bush ahead, and polls with more democrats typically have Kerry ahead or even. So, it doesn't take a rocket scientist to figure out that the election really will be based on turnout, which apparently is very difficult to determine, but historically favors Democrats. I think this election will look very similar to the the 2000 one, which means, barring any crap that could potentially happen in Florida (or Ohio), Kerry will win. Again, it's going to be ridiculously close though.

Jeff, I completely understood and I'm agreeing with you. It doesn't make sense to vote for someone who isn't electable. I'm not going to vote for me - I promise. I'm going to vote for Kerry.

CBS: GWB LEADS KERRY 51-42 AMONG LVS; 49-41 AMONG REGISTERED VOTERS...

http://www.cbsnews.com/stories/2004/09/23/opinion/polls/main645322.shtml

This must be that Kerry surge Bel was expecting. I hear CBS is inherently biased against Rep Presidents; I wonder if they fudged these numbers to make up for all their GWB animus.

CBS
Sep 23, 2004

Poll: Bush Keeps Lead Over Kerry

NEW YORK, Sept. 23, 2004: Americans are eager to see President George W. Bush and his Democratic challenger John Kerry finally debate head-to-head, and one in four voters says their decision could hinge on what they see next Thursday. For those voters, the debates could be decisive: many of them are uneasy about the President's performance and the direction of the country, but also have reservations about John Kerry.

The nationwide vote for President remains largely as it was last week: President Bush holds an eight-point lead over John Kerry in this poll. Last week, he was up by nine points. Among likely voters -- those deemed the most probable to turn out in November -- the President holds a similar 51 percent to 42 percent lead.

PRESIDENTIAL HORSERACE
(Registered Voters)

Bush-Cheney
Now
49%
Last week
50%

Kerry-Edwards
Now
41%
Last week
41%

Nader-Camejo
Now
2%
Last week
3%

Its the Same Old Song..

But Rove is singing it early this time..

Seems it is his MO to start up an inevitabilty spin before election day..last time he did it after the debates and reached his crescendo of non-sense the week before the election when he poured millions of dollars and 3-4 last minute candidate visits proclaiming that Bush would take CA and the election with nearly 400 EV's.

Of course, left with just a little bit of egg on his face in the event, he then concocted the tale which has morphed into 2004 strategery that Bush left 4 million evangelical votes on the table..,.

His comments today do not square with ARG's poll ...

In fact, Ihink given Rovian tactics, this is an admission of weakness:

"Rove told the newspaper that the swing states of Arkansas, Louisiana, Missouri, North Carolina and Virginia are already in the Bush camp"

ARG
LA Bush +8
AR Bush +3
MO Bush +6
NC Bush+6
VA Bush +6

All within MOE and since when were NC, LA, VA "swing states"?


Look carefully at what Karl Rove says...and contrast the desperation of recent statements of Bush on IraQ

If the panic has lifted from Kamp Kerry, it has settled on Baghdad Bush

Bush knows he's losing, which is why he's pulling out all the stops already. He drags in the CIA puppet from Iraq, he uses the SPR to feed big oil and lower prices a little, and he uses fear again.

OOOOoooooo. Cat Stevens!

They've nailed him and Martha Stewart. Great.

Kerry ahead in polls by October 3rd.

Re: People voting against their interest and "What's the matter with kansas"

I came to the same conclusion as this book a long time ago. It reminds me of this old lawyer joke that a fellow law student once told me back in DC. When you asked an engineer what's 2+2, he will tell you 4, when you ask a lawyer he will say "whatever you want it to be." A few disturbing things about these polls to me, and why I think Kerry is having problems (although they appear to be subsiding, the fact they happened at all makes me wonder). It is Orwellian. Where left is right, and right is left, literally in this political year. 1) I would like to know how many people supporting Bush believe there is some "secret" evidence that Sadam Hussein was behind 09/11 (I heard a guy on the radio mention this, and I know a few Republicans who have said this to me and polls bear out this belief). 2) How much of the polling is really based on images of what people think Republicans and Democrats are v. what the two candidates have actually said and done? (I think there is clearly a disconnect) 3) I am curious about how well do most people understand the legislative process? (namely the waffling label stuck not just b/c of the press, but b/c people I don't think understand how bills are made into law, and, therefore, it's an ignorance easily manipulated.). 4) I would like to see a poll that does not just ask people the horserace question of who is up and who is down, but ask them some deeper questions about their understanding of the world- ie, can they say which countries border Iraq, can they say who represents them in Congress, all kinds of questions that give a real sense not just of who they are voting for but what their understanding of who they are voting for is and their understanding of the issues outside of soundbites? This kind of polling is probably impossible beyond superficial issues questionaires but I think it would be eye opening about what the polls are really saying about the electorate- which I think is the central problem in polling.

The GOP trolls must be getting nervous. The are so desperate to spin away the obvious fact that Bush's so-called bounce is now dead that they are even stooping to use CBS polling results. Funny how the GOP thinks everything about CBS is biased and unreliable and a tool of the devil, more or less, except for when its polling shows Bush ahead.

I put as much stock in the CBS poll as I do in Gallup - both are jokes.

Make no mistake, the Republicans are getting desperate ...

I just looked at a time series chart of state poll based EV prepared from Dr. Sam Wang's Meta Analysis and I believe it shows trends rather clearly and leads to some key inferences.

Chart from the Princeton State Poll EV "Meta-Analysis"....

Kerry Median Expected EV's http://synapse.princeton.edu/~sam/pollcalc.html

1 .It is hard to say what ended the obvous Kerry bounce but notice the time of the first SBVT ad to the drop...two weeks lapsed during which time the press reverberated the charges .

2..note too that the Bush "bounce" or most of it seems - based on state polling data remember - to have occured BEFORE the RNC start (also not shown).

3. SBVT or negatives from Bush/ the press clearly brought Kerry down..not Bush's convention....

4. I don't think Karl Rove is anything near the Svengali he is made out to be but no getting around it, when it comes to garbage, he is the master..not only did the negative campaign work, it fit well with the RNC which reinforced the pre-esisting trend

5. Happily KERRY on the attack for now..you can see that too. Unfortunately I just read some of the internals in the Carville poll..IraQ is good still but the electorate is more polarized than it was two weeks back if memory serves.

SJ...
If you are going to put your faith in an outler, I should think you could do better than dismissing ARG/Zogby as Demo Polls while not mentioning DC or MSNBC or the fact that the ARG poll's sample size is 30K..

That's a short list of errors because if you are going to defend the outlier, then talk about the outliers not those within the confidence band.

Make sense?

"

Ok Patriot, You chastise me for referring to CBS; But how about AP:

GWB 52, KERRY 45; GWB JOB APPROVAL 54%

http://www.macon.com/mld/macon/news/politics/9747511.htm

Associated Press
Fri, Sep. 24, 2004

AP Poll: Bush Solidifies Support Among Men

WILL LESTER
Associated Press

WASHINGTON - President Bush solidified his advantage among men during the last month and holds his highest ratings since January on job performance, the economy and Iraq, according to an Associated Press poll.

Bush has a 7-point lead over Sen. John Kerry - 52 percent to 45 percent among likely voters - in the AP-Ipsos survey less than six weeks before the Nov. 2 election. Independent Ralph Nader was backed by 1 percent.

The president held the advantage despite increasing violence in Iraq and a week of attacks on his Iraq policy by an increasingly combative Kerry, the Democratic presidential nominee.

"We took a lead after our convention and the lead has held," said Matthew Dowd, a senior Bush campaign strategist. Bush has a slight lead in some polls, and is running even in others.

Since the Republican convention, Bush's job approval is up, 54 percent among likely voters, and just over half of them approve of his handling of the economy and Iraq. His approval in all three areas is as high as it's been all year in the polling conducted for the AP by Ipsos-Public Affairs.

Dowd wasn't assuming a Bush advantage would hold up through the election, however.

"The way we're approaching this, it will be a very close election," he said, adding that turning out GOP voters will be crucial.

The Kerry campaign focused its message in recent days on growing problems in Iraq as Bush has talked about making steady progress there.

"If you look at all the recent polls, this race is headed back to even," said Mark Mellman, a Kerry campaign pollster. He said Kerry still has time to "make the case" with voters.

"What happens in the next 40 days will be vastly more important than anything that's happened in the last six months," Mellman said. "On Election Day, there will be 10 percent who said they made their decision in the last two or three days."

In the 2000 presidential election, three in 10 voters said they made up their minds in the last month, including one in 10 who did so in the final three days, according to exit polls. The undecided group is believed to be smaller this year.

With time running out, Kerry has much important work to do in his campaign, the AP-Ipsos poll suggested.

Bush holds a 17-point lead among men. And Bush and Kerry are tied among women, a traditionally Democratic group that now favors Bush on protecting the country.

Democrat Al Gore won the women's vote by 11 percentage points in 2000, while Bush won men by a similar margin.

Betsy Bodenhamer, a 33-year-old teacher's aide and mother of two from Galesburg, Ill., says she has always voted for Democrats in recent presidential elections. This year, she's leaning toward Bush.

"I think if Kerry gets elected, he's going to pull everybody out of Iraq and they'll have to fend for themselves," she said. "Situations like 9/11 will happen again and again."

But Iraq carries risks for Bush as well.

People are about evenly divided on Bush's handling of Iraq - not that strong a rating, but better than in June when just over four in 10 approved, according to AP polling.

One thing that is crystal clear to me is that the African-American community in America literally holds the keys to 1600 Pennsylvania Avenue. The election is close enough in all of the battleground states that A/A turnout is crucial everywhere. I'm not making the presumption that African-Americans will vote Democratic, but I am looking at historical trends and where the votes are.

It's clear to me that a major commitment of resources to address this reality of arithmetic over the next month is crucial.

I trust that others are looking at the same numbers.

Being assertive about enforcing voting rights to prevent the 2000 fiasco for minority voters is one absolute.

Unless there are artificial means to hold down the minority vote in November; I believe that minority voters will tell the Bushies "THANKS FOR NOTHING" in overwhelming numbers when it counts.

In the latest CNBC poll people were asked what the most important issues were to them. 45% said economy/domestic issues. 45% said terror/Iraq. If the economy is your most important issue, you're gonna vote Kerry. So if Kerry can make even a little dent in those people who are focused on Iraq and terror then he's got a good chance.

Regarding the "orwellian" talk...I see where you are coming from but the only reason that spin from Rove and the Bushies works is that people are too stupid to see through it.

I think there should be some sort of "political literacy" test like the one mentioned above that would determine whether someone is eligible to vote. Now, that's probably not realistic because it brings back too many memories of bogus pre-Civil Rights era discriminatory poll testing. But if you did have such a test--formulated buy a national bi-partisan committee--it would screen out some of the morons who think Saddam Hussein was behind 9/11, for example. It would also force people to think more about politics and get to at least a rudimentary state of understanding about the world if they want to retain their voting rights. When I'm elected president, I'm gonna make this happen.

The New Marist poll is out. Bush leads 47-45 among registered voters, 50-44 among likely voters.

AP is out with Bush up 52-45 among likely voters. I don't know the registered voter totals.

Rasmussen shows things narrowing in today's poll. It had been 48.8-44.8 Bush a couple days ago. Now it's 47.4-46.5.

The polls tell us different things. The main message is that it's close.

And watch for SJ when he talks about "all" polls. On multiple occasions he's used that word when it in fact doesn't apply.