Battleground States 4pm Exit Polls: Prob Kerry>50%: 99.96% (he got 48.87%)

[TruthIsAll]

In the exit polls for 18 Battleground states taken around 4pm,
Kerry's unweighted average was 50.67% vs Bush 48.20%.

Two (2) states (CO and MO) moved in Kerry's direction in the
final vote (each by 1%).

In four (4) states the exit polls were EXACTLY right.

Twelve (12) states moved in Bush's direction in the final
vote.

Kerry had a  99.96% probability of getting at least 50% of
the vote.
He ended up with 48.87%.

The odds were not with John that night.
Just bad luck, I guess.

Probability of Kerry majority:
=NORMDIST(50.67/(50.67+48.2),0.5,0.0037,TRUE)

		Exit Poll 4pm			Actual Results			
		Kerry	Bush	Diff	Kerry	Bush	Diff	BushGain
1	AZ	45	55	-10	45	55	-10	0
2	LA	43	57	-14	42	57	-15	1
3	MI	51	48	3	51	48	3	0
4	IA	49	49	0	49	50	-1	1
5	NM	50	48	2	50	50	0	2
6	ME	55	44	11	53	45	8	3
7	NV	48	49	-1	48	51	-3	2
8	AR	45	54	-9	45	54	-9	0
9	MO 	46	54	-8	46	53	-7	-1
10	IL	55	44	11	55	44	11	0
11	WI	52	47	5	50	49	1	4
12	PA	53	46	7	51	49	2	5
13	OH	51	49	2	49	51	-2	4
14	FL	50	49	1	47	52	-5	6
15	MN	54	44	10	51	48	3	7
16	NH	58	41	17	50	49	1	16
17	NC	48	52	-4	43	56	-13	9
18	CO	46	53	-7	46	52	-6	-1
								
average	       50.67	48.20	2.47	48.87	50.20	-1.33	3.80
								
Prob = 99.963%						
assume 3% Moe						
								
n= sample size = 18000						
					
MoE=1.96/(2*sqrt(n))						
Std=.5/sqrt(n)						

MoE	0.0073			
Std	0.0037					


 
  

[TruthIsAll]

1. am kick

tia

[bornskeptic]

2. Slate posted some some later numbers.

You should check these out.

http://www.slate.com/Default.aspx?id=2109053&

[bagnana]

7. looking at those numbers makes me depressed.

I remember how happy I was on November 2!

[foo_bar]

3. why would 18 polls in different states contribute to the same sample?

Aren't you confusing independent statistical populations with the national population?

[Nederland]

6. Good Question (nt)

[DaveinMD]

4. as I've said

in the precinct I was working, turnout was very low in the evening and very high in the morning. This was a heavily Dem precinct. This could be what was distorting the exit polls.

[Awsi Dooger]

8. My dad reported the same thing; no doubt the cause for bad exit polls

He was appointed to precinct manager one day before the election, with two weeks total experience in campaigns, when the previous manager suddenly quit. Last night I finally got his full rundown of last Tuesday. This is in a Democratic precinct in Miami. When my dad arrived at 5:45 AM, he was stunned there was already a very long line. But at 5 and 6 PM, he says he knew Kerry was finished because almost no one was there. My father went home and to sleep at 9:30 PM, without even watching the returns.

[ISUGRADIA]

18. I had a similar experience

It was great in the morning, people lined up outside the door from poll opening 7 AM to 9 AM, steady in the day. 5 PM to 8 PM a trickle, one voter at 7:55, closed at 8 PM with no one waiting to vote.

I heard from the chair this was very odd because the evening should be busy. I checked and turnout was about the same as 2000, no big increase.

[Nederland]

5. Gender Ratio

What was the Male/Female ratio for these exit polls?

[catbert836]

9. I felt so good about 11/2

which is why I'm convinced that we'll be living in a fascist country soon.

[ArkDem]

10. Let's see. No Democrat since Jimmy Carter had gotten 50%

of the vote yet Kerry had a 99% probability. I love the new math.

[RafterMan]

19. Carter got 49.9%

Close, but not the magic number.

[tritsofme]

20. Nope, he got 50.08%

http://www.uselectionatlas.org/USPRESIDENT/national.php?year=1976

[Awsi Dooger]

21. I have seen it listed both ways

With several different verions of how many votes Carter and Ford got, thousands of votes apart. I always thought Carter received 49.9%, but most of the online resources favor your number. Dave Leip is generally the source I use, at uselectionatlas.org, the site you linked.

Here is one source that prefers the 49.9% figure: http://www.americanpresident.org/history/jimmycarter/biography/campaignsandelections.common.shtml

on edit: I looked at the discrepancies a bit further. When counting just the minor candidates who were listed on many ballots across the country, they totalled roughly 1,550,000-1,573,000 depending on which source you use. That would definitely give Carter above 50%, since his margin over Ford was closer to 1.9 million, if I remember correctly. However, many states probably had other candidates on the ballot, along with the opportunity for write-ins. For example, I saw one listing of Ronald Reagan at 1260 votes, while most places don't list him at all. I doubt the write-ins and extremely minor candidates could have totalled 300,000+, but at 4 AM that's all I'm willing to contribute or guess.

[Awsi Dooger]

22. My memory was wrong; Leip has it closer to a 1,683,000 difference

Between Carter and Ford. That would require just over 100,000 additional votes among write-ins and very minor candidates to push Carter below 50%.

[YouLiedTheyDied]

11. Let it go

The exit polls were farqued up! Even polls taken at 4:00 would not reflect the overwhelming number of thug voters that turn out after work and after dinner. I've been a poll inspector for a decade in every major city in NYS. I know how the demographics vote. Early in the day through mid-afternoon is overwhelmingly our people... After that the tide switches to thugs. This is why our ward leaders come in around 3:00 to get a list of which dems have voted and if turnout is close around 7:00, they start working the dems who haven't voted to the point of literally dragging them to the polls. Nothing wrong with that, they should have been voting anyway. But the fact is the votes come in waves and at least half of the thug vote comes after work. We're starting to look pretty stupid pushing all of these 99.9% probability arguments, its looking like sour grapes and we're going to take further losses in two years if this is what we are trumpeting!

[Awsi Dooger]

12. You are one of the best new posters on DU

Or should I say, best low-post-count DUers. Maybe you've been registered for a while. I've learned quite a bit. Keep them coming.

[bagnana]

13. I've heard this repeatedly and yet. . .

if this is the case then wouldn't the exit polls always be incorrect? Was this the first year that we had access to early exit poll results? Because this is the first year that I can remember things turning around so dramatically. It SUCKED!!!! So if this is the general flow of elections why isn't it generally noted?

[tritsofme]

14. Before the internet

almost no one one outside of the media and the campaigns would have access to early exit poll data.

[bagnana]

15. oh.

I didn't remember getting such early info. in previous years.

[tritsofme]

16. I think Drudge might have leaked the data in 2000

and 2002, and this year a lot of blogs got into the act.

Releasing early exit poll data is very irresponsible while voting is still going on.

[Awsi Dooger]

17. I posted a warning thread regarding early exit polls last Monday night

In this forum. It was titled something like, "Warning; in 2002 the early afternoon "insider numbers" were terrible."

I included at least 4 specific examples from election day 2002, all with early exit poll numbers that absurdly overstated Democratic numbers. In 2002, DUers were referencing Drudge and Rich Lowery of the National Review, who were each releasing early numbers online that turned out to be way off.

[TruthIsAll]

23. Your way off. I prove it right here.

http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=104x2645318

The more you deny the obvious, the more irrelevant you become.

Even that toe-sucker Dick Morris says the exit polls are always right. He says there was foul play.

And so will everyone else, real soon now. Keith Obermann broke the ice last night.

Soon you will be left out there all alone, spinning for Rove and BushCo.

They cheated. This time it will be proven. Totally.

[ArkDem]

25. Your is a pronoun.

[Awsi Dooger]

26. You are beyond irrelevant when you insist on embracing phony numbers

Or should I say, only numbers that are most favorable to Democrats. That was pathetic in 2002, in fact embarrasing for someone obviously as skilled in math as you are. Your use of only the most favorable polls from 2002 to make a case we were robbed in 2002 was among the most assinine and misleading arguments in DU history. When there is a late national tide, as caught by a USA Today headline on Monday before the election, of course state polls conducted days earlier are no longer valid.

Here, merely off the top of my head I'll provide the examples: Cleland way ahead in Georgia (late polls were tied, with Chambliss all the momentum), Mondale far ahead in Minnesota (final polls showed either guy ahead), Shaheen leading Sonunu by 6 in New Hampshire (late polls has Sonnunu ahead), Kirk only down 1 in Texas (that was so atypical from the normal estimation of down high single digits it was even absurd by your standards to include it), Strickland leading big in Colorado (yeah, we were blowing out an incumbent in Colorado).

This year I cautioned you and others time and again that state polls are unreliable, that there is a logical and predictable relationship between national popular vote margin and individual states. You were too blind or partisan to agree, and kept including state polls in your model that made zero sense in regard to national preference. No chance Ohio was ever going to tilt significantly in our favor from the nation itself.

BTW, I noticed you did not address the point I (correctly) made in the post you replied to, that early exit polls in 2002 were similarly garbage. I inserted 4 specific examples, and this was a night BEFORE this year's election, without knowing which way the early exit poll numbers would break this year.

Republicans obviously cheat, but in traditional ways like vote suppression and intimidation. I believe the machines are accurate and will eventually help us, particularly in large urban areas. I have continuously argued for mandatory paper trails and an automatic check with machine results regardless of margin. Don't you dare imply I'm on their side. Meanwhile, while you guys are fruitlessly trying to overturn an election we LOST, I'm still paying daily attention to one we can win, namely the Washington governors race. No doubt you can't even name the candidates, and won't be able to unless some pre-election or exit poll differs significantly from the result and we lose. Hint: the Democrat was expected to win. Her own campaign manager admits their final polling caught an unexpected Republican wave. I'll be more than happy to link that quote if you post we were cheated.

[Name removed]

27. Deleted message

Message removed by moderator. Click here to review the message board rules.

[Bridget Burke]

24. Has any of this been published?

You've posted this same theory in several threads. If it's true, surely there's further evidence--outside your own experience, that is.