« Swing State Polls: Kerry Running Strong | Main | WSJ/Zogby Poll: Kerry Ahead or Tied in All Battleground States »

New AP--Ipsos Poll: Kerry, Bush Tied

John Kerry and George Bush are tied at 47 percent of nation-wide RV's on the eve of the second presidential debate, according to an AP-Ipsos Public Affairs Poll conducted 10/4-6. (Kerry leads Bush 50-46 percent among LV's)

According to AP's Ron Fournier, "Fewer voters than a month ago believe Bush is the best man to protect the country and fight the war.

"The AP-Ipsos Public Affairs poll, completed on the eve of the second presidential debate, charted a reversal from a month ago, when the Republican incumbent had the momentum and a minuscule lead. Since then, bloodshed increased in Iraq, Kerry sharpened his attacks and Bush stumbled in their initial debate.

"Nearly three-fourths of likely voters said they had watched or listened to the first presidential debate last week, according to the poll. Only 8 percent came away with a more favorable view of Bush while 39 percent said they felt better about Kerry."


I've asked this question elsewhere and never got an answer: has there ever been a time when an incumbent trailed in a serious presidential poll this close to an election and came back to win? I suppose Truman would be one case, though apparently, back then, pollsters stopped surveying right after Labor Day, believing opinion was solidified by then. It strikes me as a very ominous sign for Bush.

It's hard to find any polls at this point that give Bush 50% or better; the only distinction among the varoius polls is how low Bush's number goes (somewhere from 45 to 49), and how much of the opposition has so far declared for Kerry. Zogby, for instance, has it Bush 46/Kerry 44, but his profile of the undecided makes it clear they're most Kerry votes waiting to happen.

Meantime, the reports of new registrations read like a DNC fantasy. If these numbers translate into real turnout gains in November, we could be looking at something extraordinary.

Okay, note to self: turn off the excitement meter and hunker down. 26 grueling days to go.

Are we supposed to be looking at poll results from likely voters or not? Isn't Bush over 50% in any poll bad news....

Has anyone created an algorithm that translates a popluar vote into the most probable electoral vote? Is that even possible?

Well, this is certainly encouraging news for Kerry, although I don't know what to think about the reason(s) for it. Too late for a post-debate bounce from last week. Is there any such thing as a pre-debate bounce?

This has been a week from hell for the GOP on Iraq. Meanwhile, Bush and Cheney continue to insist that Saddam might not have had weapons or the means to develop them, but he WANTED them, so the war was the right thing to do. And, by the way, everything's goin' fine. Maybe LVs and RVs across the country are starting to think that the president and vice-president have actually lost their minds.

I'll tell you what, though. This polling roller coaster is starting to drive me crazy. November 2 can't get here fast enough. I can't take much more of this.

I think Bush is very close to the "tipping point" where things will quickly start to unravel. If Kerry crushes Bush in Friday's debate, which i believe he will, the momentum might be unstoppable. As GWB's platform is built upon a bed of lies, his campaign might crash harder than all of us suspect.

using this poll and all other nationals polls via pollingreport.com

Difference of 1.16%
bush 47.6
kerry 46.5

not using polls if they are greater than 2 STDEVs from mean

bush 47.33
kerry 46.66

its a 0.66% difference. if we expect 110,000,000 voters than thats a difference of just 730,000 people

Waiting for some comments on the WSJ-Zogby Battleground Poll completed yesterday:


Adding in all leaners within margin of error gives Kerry the win with 322 to Bush's 216. Removing the states within the margin of error puts it Kerry: 243 electoral votes to Bush: 189. This appears to be the largest number of electoral votes outside of the margin of error for either candidate since Kerry's 252 on 7/12. (Other previous bests are 235 for Kerry on 8/2 to Bush's 225 on 8/23.)

WSJ's analysis basically tries to undermine some of the poll they sponsored. Interesting.


There's no good news for Bush in this poll. None.

This is going to be a breakout year for voting. High registration, high motivation (on both sides) and improved GOTV efforts will result in the highest participation we have seen.

Polls aren't able to track the new voters that only have cell phones, or don't meet the likely voter filter. Frankly, that filter is clogged, and needs to be cleaned. This year is nog going to conform to the model of previous years.

It's close now, but the momentum has shifted. Be careful of pitfalls, and watch out for an Octuba surprise.

Bush basically has done a terrible job in the last 4 years, so he doesn't have much to run on. This week his best reason for invading Iraq was blown out of the water. His post invasion strategy was shown to be flawed. And he showed his ugly scowl to over 60,000,000 on tv.

The Los Angeles Times had an editorial on Thursday, October 7, 2004 saying
"Is He a Dope?
Although neither group likes to say so, some Americans who support President Bush and many who don't support him have concluded over four years that he may not be very bright. This suspicion was not allayed by Bush's answers in the first presidential debate a week ago."

TRUMAN VERSUS DEWEY REDUX: AP PROJECTS GWB THE WINNER...It appears AP got carried away earlier today. Toooo Funny.


Oct 7, 2004

AP: Bush wins election, Republicans keep House/Senate

The AP earlier today accidentally posted an election story that claimed "President Bush has won re-election as president by a 47 percent to 43 percent margin in the popular vote" and Bush "has won 324 electoral votes in 33 states. He is leading in 4 states for a total of 43 more electoral votes". The Story, printed on ABC's Green Bay Wisconsin affiliate's website Wbay.com is apparently an Election day test that was accidentally place on the world wide web.

Update: Here is a screenshot of the story http://www.stopdemocrats.com/images/wbay.jpg. The story had been posted on AP affiliate sites in Wisconsin, Delaware, Illinois, Arkansas, and Arizona.

Announcement: We will be live blogging tommorrow's crucial Presidential debate.

The 7 day rolling averages of 5 key battleground states are:
Florida, Bush 52%-45% (This is a 1 point Bush gain, and has placed Florida in the "Leans Bush" column by Rasmussen for his EV count)
Michigan, Kerry 49%-46% (This breaks a week long tie, and is a 3 point Kerry gain from yesterday)
Ohio, Bush 50%-45% (This is a 3 point Bush gain from yesterday)
Minnesota, Kerry 48%-45% (This in unchanged from yesterday)
Pennsylvania, Bush 47%-46% (This is a 1 point Bush gain from yesterday)

Flash: Newly discovered audio files from a 2002 CNN interview reveal John Edwards thought "Iraq and Saddam Hussein present the most serious and most imminent threat {to our country}"

More pro Kerry polls for Iowa, Minnesota, Florida, Wisconsin and New Mexico...




Thoughtful comment Ramdan and I would add that it all depends on the distribution of a relatively small number of voters in a small number of swing states. A nail biter, at this point anyway.
I certainly hope the tightness of the race is giving Nader voters pause.

Also, demtom, back in 1948 they basically didn't poll undecideds, and many of the national polls vastly overrepresented the pro-Dewey northeast.

I know this blog is for polling, but below is an interesting article that I think sums up the campaign from here on out:


Things are looking better for Kerry in the polls but he still shows quite a bit behind in the electorial college. It seems like if he is nearly tied in the polls he should be tied in the college.

Actually, Kerry's further ahead in the electoral college than he is in the general population.

I'm not sure why, but it takes time for changes in national polling to be reflected in polling from individual states. The states usually lag the national trends. However, electoral-vote.com, for what it's worth, has it Kerry 253 Bush 264. Not bad.

On the Truman-Dewey race in 1948.

This is a classic that has long been taught in survey research courses -- and there are some interesting points that suggest things to consider this year.

First of all, the last Gallup poll was conducted in mid October. It did not pick up Harry Truman's passionate late campaign by rail, and it did not pick up the near collapse of the Wallace campaign in the last weeks. Remember, 1948 was a 4 way campaign, Truman and Dewey, plus Strom Thurmon on the Dixiecrat ticket, and Henry Wallace running as a progressive.

But the real polling era was Gallup's -- they had not considered the need to adjust their polling results to the vast demographic changes that were the result of World War II. In many respects, they still operated off demographics from the 1940 census -- and the War had changed lots of things.

In particular, Northern Industrial States had many precincts that were 3-4 times as populus post war as pre-war. But housing was scarse, and people were living doubled and tripled up. THIS WAS PARTICULARLY TRUE IN THE BLACK COMMUNITY. With the labor movement putting massive effort into registeration and GOTV that year -- Industrial worker precincts were way underpolled given their population.

But it was really the Black Community that made the critical difference. In fact, Gallup did not really poll blacks, and it employed no black interviewers, and apparently they just assumed Blacks were Republican. But that wasn't true any more. Blacks who had migrated from the South -- from can't vote states to places like illinois and PA where the CIO would actually take them down to register after work -- fully understood Truman had started to integrate the services, supported Fair Employment Practices, had voted in the Senate for the anti-Lynching laws -- and most important, had the courtesy to ask for votes. It's the moment of the "big switch" which had been underway for some years. Blacks turned out big time in 1948, and provided the margin in Industrial States that put them over the top for Truman.

In many ways this election could be somewhat parallel to 48 in that the technology change from land lines to cell phones could be systematicly missing a significant segment of the electorate. Likewise, I am not certain we yet know how the GOTV strategies of this year are going to work.

the correct link for WSJ's battleground poll results page is this one: