Latest Rasmussen 3-day average: Bush 46%, Kerry 45%

rasmussenreports.com ^

[WI Conservative 4 Bush]

Bush 46% Kerry 45% Election 2004 Presidential Ballot

Bush 46% Kerry 45% Other 3% Not Sure 6% RasmussenReports.com

--------------------------------------------------------------------------------

Rasmussen Reports Home

Thursday April 15, 2004--The latest Rasmussen Reports Presidential Tracking Poll shows President George W. Bush with 46% of the vote and Senator John F. Kerry with 45%. Roughly half the survey interviews were completed before the President's Press Conference on Tuesday night.

Single night tracking results for Wednesday show Bush with 46% and Kerry with 43%. This is the only night of polling since the Press Conference. Single night tracking data should be viewed with caution due to relatively small sample sizes and other factors. Tomorrow's sample will include two days since the Press Conference and provide a more solid indication of whether there is any real movement.

[WI Conservative 4 Bush]

Latest from Rasmussen. It's a long road to November...

[DM1]

I really hope that W can pull this off
@#$% Iraq is killing him thankfully Kerry is a stiff so hasnt capitalized on it as much as someone like Clinton could but if this keeps up we are going to lose this one save a miracle

[Kay]

Not surprised. President Bush did well.

[TheErnFormerlyKnownAsBig]

I can't remember, is this poll registered or likely voters.

[not-alone]

it still is good news

[WI Conservative 4 Bush]

It's likely voters, 500 per night through an automatic phone system.

As has been discussed at length on FR, Rasmussen should not be taken very seriously as far as the raw data. However, movement in this poll is what I find attractive...

[KellyAdmirer]

To the extent this kind of thing is useful at all at this point, I think using only 3 days to calculate an average is silly (I know this is considered somewhat standard, but this poll jumps around way too much). Too much fluctuation due to a single day of polling. Five or six day averages would smooth out the fluctuations a bit and make this relatively more meaningful. If anyone wants instant reactions, you can always look at the "latest polling night" data.

[Patrick1]

Rasmussen who completely blew it in 2000 has a long way to go to get his credibility back, and this daily tracking poll of Kerry up 6% on Tuesday and down 2% on Thursday ain't helping.

[GWB00]

Does anyone have some historical information on previous Presidential match-ups? I understand that Dukakis was ahead of GHB Sr. and Mondull was ahead of Reagan at this time in the election cycle.

[Patrick1]

Dukakis lead GHWB coming out of the Dem convention by double digits in 1988 and Mondale lead Reagan in February, but in reality that was media wishful thinking. Reagan had won re-election the day he announced in January 1984.

[TomEwall]

O'Reilly keeps quoting the one day where Kerry was ahead by 6 points as if it were etched in stone.

[TomEwall]

I think you're right, about the movement. So far it looks like a random walk model. They seem to be consistently understating Bush's numbers, probably due to not including Nader.

[EternalVigilance]

To President George W. Bush on his visit to my homestate of Iowa today:

Mr. President, carry on
Though the battle may be long
Be brave
Be strong

Do not heed the voices
Of those who would appease
Defeat is the only place
That would lead

Stay the course, the one you set
It hasn't failed yet
Stay on the offense
It's our best bet

Fight on, in memory of those who have given their all
They willingly answered their nation's call
If you do this
Those who hate us will certainly fall

You have our prayers, we stand with you
Together we'll do what we have to do
To guarantee that our children
Will enjoy liberty, too

EV

[Patrick1]

Indeed. So many panic over these daily polls, most of which are useless when you look at the internals. Newsweek is the classic one for this kind of junk news making.

The fact is, as I wrote on FreeRepublic a while back, JFK would have to run a near perfect campaign to get enough electoral votes to win election. My prediction remains that GW will beat JFK by five to six percentage points and have a very solid electoral college victory.

[RetiredArmy]

This poll has more ups and downs than the Rocky Mountains.

[Owen]

Guys,

No one can know the right sampling mix. No one. How would you do it? Would you say . . . okay, I want to be sure I sample 50/50 women/men. No wait, I want it to be not 50/50, but rather the % gender mix that voted in 2000. Or no, maybe I want it to be the mix from 2002. Or, maybe I want to say . . . more men than women are attending political rallies this year so I'll give them an edge.

Okay, so I have my gender mix. Now, what about zip codes? I want to oversample the zip codes where high % of turnout took place in 2000. Or wait, should I use 2002? Well, one or the other, but wait, hasn't there been some demographic migration from zipcodes since 2000? Or 2002? Maybe I need to adjust that.

Now, what about party affiliation? What mix should I use there?

Okay, there. Done. Now, here are the results.

[evilC]

With a 4% Margin Of Error, I guess that means with a 95% certainty (assumed usual confidence level) either Bush or Kerry will win.

I could have told you that for free!

[evilC]

Oops! Should be 3% MOE.

[CommerceComet]

Although you point out how hard it is to get the proper mix of several factors, you left out one great complication. All the polls are trying to predict what will happen in 2004. Although historical measures are probably the best predictor, there is no assurance that they will be truly predictive. 2004 is unique and the turnout may be influenced by some factor which was not important in the past.

[Owen]

Yeah, finding the right mix of 2004 is guesswork, and what astounds me is just how much money is being paid how may different pollsters to grope around in that guesswork.

Lot of money to be made polling. Accuracy is not important. Sort of like stock market analysts.