It happens all the time.

[TruthIsAll]

Initial Exit Poll        Final Vote	          Deviation	
	Dem	Rep		Dem	Rep		Dem	Rep
1988	50.3	49.7		46	54		-4.3	4.3
1992	46	33.2		43	38		-3	4.8
1996	52.2	37.5		49	41		-3.2	3.5
2000	48.5	46.2		48.4	47.9		-0.1	1.7
2004	50.74	47.93		48.28	50.73		-2.46	2.8
								
Average	49.55	42.91		46.94	46.33		-2.61	3.42

The initial exit poll is massaged to match the recorded,
official vote tallies.

Does that mean the initial exit poll was wrong?
When we all know that millions of mostly Democratic votes were
never counted in the first place?
So how can the final, "weighted" poll be right?

Am I missing something here?

Consider this:
If a blackjack dealer were to take two ACES out of the deck
before dealing the first hand, would you sit down to play?

[BlueEyedSon]

1. Did you ever do new reg by party vs new voter by exit poll vote (party)?

[Aussie_expat]

2. Black Jack answer..

"If a blackjack dealer were to take two ACES out of the deck before dealing the first hand, would you sit down to play?"

Yes, because everyone would be playing without two aces.

[TruthIsAll]

3. You obviously never learned how to play scientific blackjack.

The fewer high-value cards (10, queen, jack, king, ace) left in the deck, the worse it is for the player. And vice-versa.

I expect that answer from you, since you don't understand basic probability, whether it's the percentage of "rich" (high) cards left in the BJ deck, or the percentage of Democratic voters left to be counted after millions of ballots have been rejected due to "spoilage".

[Wilms]

4. kick n/t

[WI Independent]

5. The raw numbers are not based on a random sampling

and therefore will always be inaccurate until they are weighted. To be random, every voter would have an equal probability of being polled. We know this is not true since most polling places had zero samples taken. Weighting is required to correct for the lack of randomness. Weighting is based on assumptions about the demographics of the voters that will vote. As actual results come in these assumptions can be proven wrong and adjustments required to the weighting assumptions.

[TruthIsAll]

6. I suggest you read the notes at the bottom of the WP screen

This is the "preliminary" National Exit Poll.
It's a randomly selected 13,047 sample.
With a 1.0% Margin of Error

The weightings appear very reasonable:
The Party-ID is 38 Dem/35 Rep/27 Independent.
And Kerry is the 51-48% winner for all category weightings used to calculate the vote. But this 13,047 "preliminary" poll vanished into the ether, never to be heard from - except on the Net.

Since the poll consisted of "raw" data (even though the sample was randomly selected - see the notes) it needed to be "weighted" to match the recorded vote - along with adding 613 new voters to the original sample, bringing the final total to 13,660 respondents. And thus we now have the official National Exit Poll.

All of these 613 "new" voters must have voted for Bush in order for the exit poll numbers to match the vote. Of course, the demographic weightings also had to be changed accordingly.

The preliminary 13,047 poll had a Party-ID mix (randomly-sampled) of 38 Dem/35 Repub/27 Independent. This demographic also happens to COINCIDE with the Party-ID mix for the FINAL, WEIGHTED exit polls in 1996 and 2000 (39/35/26).

SO THE PARTY-ID WEIGHTINGS HAD TO BE CHANGED TO 37/37/26 IN ORDER TO MATCH THE BUSH 2004 RECORDED VOTE.

Voila!

The final exit poll of 13660 shows that Bush won by 51-48%.

NOTE THE EDISON/MITOFSKY CAVEAT: "FINAL PERCENTAGES MAY SHIFT SLIGHTLY"

Remember the Lovin' Spoonful in the sixties?
"Do You Believe in Magic"?

HERE'S THE LINK TO THE GRAPHIC ON THE NEP/WP SITE (POSTED NOV.3):

http:
//media.washingtonpost.com/wp-srv/politics/elections/2004/graphics/exitpolls_us_110204.gif