Final State and Pre-hacked Exit Polls Match: Average Deviation = - 0.18%

[TruthIsAll]

Median deviation: -.05%
Standard deviation: 3.025%
Kerry was leading in BOTH the pre-election and exit polls.

The deviations were spread around the -.05% median:
No bias to Bush. No bias to Kerry.

The bias to Bush began with the POST-MIDNIGHT adjustmnets made to the exit polls to match the late Bush "surge".

Zogby was right on. So was Mitofsky. Despite their claims to the contrary. All we are hearing is that the Exit Polls were wrong. Well, this proves otherwise.

If the published polls were right, then so were the exit polls. Kerry WON.




State Deviation
1 VT -8.01
2 AK -6.02
3 NH -5.40
4 UT -4.69
5 DE -4.28
6 NY -3.63
7 OK -3.54
8 DC -3.36
9 HI -3.30
10 CT -3.18
11 RI -3.13
12 SC -2.70
13 NE -2.34
14 PA -1.72
15 MD -1.33
16 NM -1.30
17 MT -1.04
18 IN -0.79
19 WA -0.78
20 IL -0.75
21 OH -0.55
22 AL -0.37
23 MN -0.33
24 MO -0.19
25 CA -0.15
26 WY -0.05
27 KY 0.05
28 NC 0.14
29 ID 0.21
30 AZ 0.37
31 CO 0.37
32 NV 0.65
33 LA 0.95
34 NJ 1.01
35 FL 1.04
36 MI 1.11
37 WI 1.18
38 ME 1.43
39 TX 1.54
40 GA 1.68
41 MS 1.91
42 OR 1.99
43 VA 2.33
44 IA 2.54
45 KS 3.14
46 WV 3.27
47 AR 3.40
48 MA 4.33
49 ND 4.89
50 SD 6.93
51 TN 7.23

[Hobbes199]

1. Why hasn't Mitofsky stood by their results?

You'd think his reputation would be on the line, and that it would outweigh any political pressure.

[Helga Scow Stern]

2. Maybe his allegiance is more important than his rep.

[Hobbes199]

3. Could be, but...

Never underestimate the power of the ego. I think he needs to get some good pressure.

[sw40fsks]

4. Which pre-election poll are you using?

Those polls were all over the place, probably the most reasonable numbers to use were those which calculated an average of a bunch of different polls, some showing JFK ahead and some showing the shrub. I was generally following this one:

http://www.realclearpolitics.com/polls.html

This is the final one from the night before the election. It seems to match the final results pretty closely.

[99Pancakes]

5. Real Clear Politics

never factored in the Undecideds, so it was not the best site to be at.
:argh:

[sw40fsks]

6. Some did, I think

I think they left that up to the individual pollsters. For example, I'm pretty sure the final Gallup poll had the undecideds allocated, didn't they?

[99Pancakes]

11. Real Clear Politics

listed no undecideds.

[99Pancakes]

12. Real Clear Politics

listed no undecideds in the battleground state polls.

[Silverhair]

18. Yes, but Gallup allocated the undecided 90% for Kerry.

Obviously, the undecideds didn't break that way.

[TruthIsAll]

8. I used a variety of polls in the Election Model.

Zogby, ARG, Research 2000, even a few from Survey USA, Strategic Vision et al.

[99Pancakes]

13. Yes, so did I. nt

[TruthIsAll]

10. Which finals? Earlier exit polls or those fixed to match the hacked votes?

tia

[newyawker99]

29. Hi sw40fsks!!

Welcome to DU!! :toast:

[BlueDog2u]

7. Kick!

Awesome work.

This needs to be distributed more widely. What are your plans?

[TruthIsAll]

9. I'll let others decide if it's useful.

tia

[bleever]

14. Zogby, electoral-vote.com, and other uber-pollsters and oracles

all knew.

Zogby, if he's going to have a commercial future to be proud of, will have to get into this.

The wrath of the oracles, however, is simply inescapable.

[bleever]

15. And let's not forget, we could tell that B* knew it too! Karen Hughes

had to sit him down late in the day and break the final bad news to him, but we had already seen him schlumping around with a pinched look for the last 36 hours before the election. They knew it was grim.

[TruthIsAll]

16. Why would they reveal that Hughes sat down with *?

Setting the stage for plasuible denial?

[bleever]

20. My take is that this was an actual leak, and that the later

photo op may have been necessary to restore the facade of confidence which was needed to offset any national surprise over the turn of the tide in B*s favor.

I'm inclined to believe that plan A was to see if they could win honestly (taking into account all the vote suppression already in place) and not risk getting caught, and to go with plan B only if needed.

Very likely B*, as well as Hughes (who has an uneasy rivalry with Rove) were not in the loop on plan B, and here's where plausible deniability comes in. But there was a glitch: Hughes took it upon herself to prepare the pResident for what was shaping up as a big loss, and the word got out before Rove's plan B starting showing results.

[Silverhair]

17. Take a look at this Mason-Dixon site.

http://www.mason-dixon.com/news_text.cfm?news_id=231

It doesn't help if you only look at the polls you like and dismiss the ones that you don't like. Before the election Mason-Dixon was showing a marked Bush lead in several of the battleground states. The link is MDs site that compares their polls to the actual results.

Edit: Typo

[goldengreek]

22. If I understand TruthIsAll correctly...

This is the result of exit polls compared with a composite of tracking polls going into the election. With different people using different methodologies you're usually going to get closer to the final result. Your link only connects us with one polling source.

Is that right, TIA?

Remember: we ARE finding concrete cases of fraud.

[Silverhair]

23. I used the one with the best long term record of accuracy.

Some of the polls that TIA uses are suspect at best. Using a known poor poll, just because you like the results of the poll is still going to give you bad answers. It is like a certain politician surrounding himself with "Yes" people.

Bush won and the sooner we face it, the sooner we can plan strategy for the next election. Spinning our wheels on super conspiracy theories doesn't get us anywhere.

Has everybody here forgotten that Bush was leading in almost all the polls, by about 2 - 4 points? When a candidate wins by a margin of 2+ points, they are pretty much sure to carry the EC. Yes, I know that the margin in OH was about 150K, but to change that 150K Kerry would have also had to have made gains in the rest of the country at the same time. The fact is that Bush showed improvement in almost EVERY state. Where he lost he lost by smaller margins, with only a couple of exceptions. That is the biggest proof that he did indeed win.

Screaming, "We were robbed!" only makes Democrats look stupid.

[goldengreek]

24. Thanks, Silverhair.

I sigh because it looks like I'll have to start digging deeper and I don't have a lot of time for that. Still, I'll rely on composites of polls because I believe they'll get closer to the truth.

But we need to face the reality that if this election wasn't stolen there's no way we'll never win again unless we have an "H. Ross Perot" window opening up again. Bush has failed at every test handed to him; he should have been a sitting duck.

[Silverhair]

27. We can win.

But first we have to ask some hard questions.

[farmbo]

19. Why not hire Zogby to do some 'after polls' in key areas?

We could have him poll high numbers in those Fla "Dixiecrat" counties to invalidate (or confirm) the Rovian explanation.

He could also poll key counties in Ohio to see whether those anomalies may have flipped the election.

Hell...I'm good for $100... If TIA signs off on it.

[goldengreek]

21. Fire at will!

We can't let these people run away from the exit polls. We're looking at several different groups using several different techniques and they're all coming to the same conclusion: FRAUD. And Bev Harris is getting to the nuts and bolts of it.

[TruthIsAll]

33. Pray to the Greek Gods of Homer.

tia

[Goldeneye]

25. How close are the final pre-election polls

usually to the exit polls? Is there any data on this?

[Zan_of_Texas]

26. Help.

What does it mean when the blue block in the bar chart is below the line, vs. when the blue block is above the line?

Could you add notes to the graph to explain EXACTLY what data are being used, hopefully with links to them.

It looks to me like there IS a fair amount of deviation in the pre-election and exit poll results (which exit polls should be specified).

And there's no examination of the "actual" election tallies here, by state, compared to either polls or exit polls?

Sorry, I'm no statistician, but I do love graphs.....

[TruthIsAll]

28. This will clear up the calculation of the state deviations.

The graph shows the percentage deviations of the Kerry 2-party state poll from the corresponding exit poll.

Her's a quick example.

Assume the final PUBLISHED poll for a state is Kerry 50%, Bush 40%.
Then Kerry's two-party percentage is 50/(50+40) = 50/90 =55.55%

Assume the corresponding state EXIT poll is Kerry 53% Bush 47%.

Then the Kerry deviation is -2.55% = 53.00% - 55.55%

[Zan_of_Texas]

30. Okay. Let's talk about math language.

It's been about two geological eras since I took statistics.

But, my recollection is, when you use words like "deviation" and "standard deviation" you are talking about a number that is derived from some serious math maneuvers, like doing deviation from a mean or something.

What you are doing would be more clearly called a "difference", because it is a simple A minus B number.

Is that right? Statistical whiz types, help me here.

[TruthIsAll]

31. The difference is the deviation: A-B. The standard deviation is..

The Standard deviation is a measure of variability around the mean.

68% of the observations fall within 1 standard deviation of the mean.
95% of the observations fall within 2 standard deviations of the mean.

This is from the Excel Help Screen:

STDEVP
Calculates standard deviation based on the entire population given as arguments. The standard deviation is a measure of how widely values are dispersed from the average value (the mean).

Syntax

STDEVP(number1,number2,...)

Number1, number2, ... are 1 to 30 number arguments corresponding to a population. You can also use a single array or a reference to an array instead of arguments separated by commas.

STDEVP assumes that its arguments are the entire population. If your data represents a sample of the population, then compute the standard deviation using STDEV.

For large sample sizes, STDEV and STDEVP return approximately equal values.

The standard deviation is calculated using the "biased" or "n" method.


Example: Given this set of 10 values, what is the standard deviation?

1 1345
2 1301
3 1368
4 1322
5 1310
6 1370
7 1318
8 1350
9 1303
10 1299

Average (mean) = 128.6

Plug the values into the Excel function = STDEVP(A2:A11)
The Standard deviation= 26.05

Approx. 68% of the values fall within +/- 26.05 of 128.6, the mean (average) of the 10 numbers.

[TruthIsAll]

32. Kicking my own post

with 100% probability