GREAT NEWS! THIS IS WHAT ZOGBY MEANT WHEN HE SAID KERRY HAD A GOOD DAY...

[TruthIsAll]

I calculated what the Kerry and Bush one-day poll numbers had
to be in order to come up with Zogby's three-day moving
average poll released today: Bush is leading 46-44, down from
48-44 yesterday.

It is not obvious that Kerry made a big move in yesterday's
polling. After all, his 3-day moving average held constant at
44.

But....Kerry outpolled Bush yesterday 49.6 - 39.1!

Zogby doesn't want you to know it.

Job security.



	TEST TO DETERMINE ZOGBY ONE-DAY POLL					
Date	bush	bush	bush	kerry	kerry	kerry
Oct	zogby	day	3dMA zogby	day	3dMA
6	46	46	46	44	44	44
7	46	46	46	45	46	45
8	45	43	45	46	48	46
9	45	46	45	46	44	46
10	44	43	44	47	49	47
11	45.2	46.6	45.2	44.6	40.8	44.6
12	45.3	46.3	45.3	44.8	44.6	44.8
13	46	45.1	46	45	49.6	45
14	48	52.6	48	44	37.8	44
15	48	46.3	48	44	44.6	44

16	46	39.1	46	44	49.6	44

[NMDemDist2]

1. wowsa n/t

[trumad]

2. After this election I'm buying you a beer

on second thought a Case!

[Mz Pip]

3. So how

long doew it take for Kerry to overtake Bush and develop a 4-6 point lead?

I need some good news. I need a good night's sleep. ;-)

MzPip
:dem:

[pa28]

4. That is great news.

The polls should take care of themselves if everybody on our side keeps hammering away and stays upbeat.

[naufragus]

5. The DNC needs to hammer the media

the story today should not be poll numbers bythe NYT and the guardian have basically said bush is a wacko. we need some pit bulls on tv

[ClassWarrior]

18. I can't do anything about pit bulls on TV.

But I did volunteer for JK/JE for 5 hours today! GET OUT AND VOLUNTEER EVERYONE!!

24.

[gasperc]

6. this will hit the moving average by tomorrow and by Tues

really show the spread, I'm certain we'll need to brace for red alerts and what other slime fest they expect to toss between now and nov.2

[gasperc]

7. but boy, how unstable look at the 14th numbers

Bush was up to 52.6% and Kerry down to 37%, and the day before Kerry polled 49%?? that would be Thursday after the debate what was up with that poll?

[TruthIsAll]

16. If Kerry ties Bush 45-45 today, he'll be leading 46.4-43.5 in

Monday's Zogby poll...

[Longhorn]

8. Why doesn't he want us to know it?

And thanks for giving us something to look forward to!

[batchdem04]

9. I don't understand...

Zogby says "The third debate is now registering among voters and Kerry had a good day, leading by two points today alone"

But you have him with a 13 point lead?

[TruthIsAll]

12. Check my numbers against Zogby's. The only way the 3-day

Zogby moving average can be replicated is if the one day totals are as shown.

I assume Zogby is taking the arithmetic average of the last 3 days polling results. If he is weighting these numbers, then my calculations would be off.

I have Kerry leading 49-39 in yesterday's polling, a 10 point lead.

[Wakeup_America]

15. I don't think you can reproduce

the daily numbers uniquely unless you know at least two of the daily numbers. Check out my numbers posted above. According to my numbers Bush-Kerry=5.4 that averages to 5.4/3=1.8 or roughly 2 points.

[gasperc]

14. that's the results of the latest 2 day poll, the 3 day ave. is 44%

however, tomorrow the 37.6% reading will be dropped from Kerry's ave and if the oct. 17th poll is anywhere as good as the 16th poll then Kerry's #'s will show a huge surge. In other words, better than 2/3 of the soft and undecideds are breaking to Kerry

[unfrigginreal]

10. TIA

There's something not quite right here. Zogby is saying that Kerry polled +2 on the day. See here:

"The third debate is now registering among voters and Kerry had a good day, leading by two points today alone. He now is behind by only two overall. He managed today to regain a good lead among 18-29 year olds (especially 18-24 year olds, who he will need on November 2 to win). He also has consolidated his lead among Hispanics and is now only yielding 12 points to Bush among Democrats.

http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=132x1093710

[batchdem04]

11. Well one thing is for certain

is that there is a 6 point difference from Bush's polling number on October 13 and October 16.

Real quick:

Oct 13 : a
Oct 14 : b
Oct 15 : c
Oct 16 : d

(a + b + c) / 3 = 48
(b + c + d) / 3 = 46

a + b + c = 144
b + c + d = 138

144 - 138 = 6

[Wakeup_America]

13. Daily numbers

TruthIsAll, How did you come up with the daily numbers? My program (based on minimizing day-day variations) produces these numbers:

Day Bush Kerry
10/6 45.7 45
10/7 47.5 43.7
10/8 42 49.4
10/9 45.7 45
10/10 44.5 46.7
10/11 44.9 43.4
10/12 45.7 44.9
10/13 47.5 46.7
10/14 50.9 40.4
10/15 45.7 44.9
10/16 41.5 46.7

The numbers basically show that Bush had a great polling Thursday right after the debate and Kerry polled well (47/42, not as good as your numbers) yesterday.

[TruthIsAll]

20. Trial and error in Excel. I varied the daily number to equal Zogby's

tracking poll numbers, which are just three-day moving averages.

Obviously, I could not calculate the moving average for the first two days of the tracking poll, so I assumed that the daily poll numbers were equal to the tracking poll numbers for those days.

If my methodology is wrong, let me know why.

[TruthIsAll]

21. Zogby is just one of 18 polls. Kerry has a 99% win probability

For you geeks who have not seen my Election Model site, I think you will enjoy the analysis and graphics:

http://www.geocities.com/electionmodel/

[Wakeup_America]

24. The first two numbers

obviously don't have to be the same as the first two moving averages! As you can see from your numbers when you do make that assumption, the future numbers will tend fluctuate wildly.

In my opinion, one reasonable way to 'estimate' the daily numbers is to minimize the overall day-to-day fluctuations so that the resulting variance will be within the MOE for the daily sample. So as you can see from my numbers

Day Bush Kerry
10/6 45.7 45
10/7 47.5 43.7
10/8 42 49.4
10/9 45.7 45
10/10 44.5 46.7
10/11 44.9 43.4
10/12 45.7 44.9
10/13 47.5 46.7
10/14 50.9 40.4
10/15 45.7 44.9
10/16 41.5 46.7

The first two numbers are not the same as the moving averages but the daily numbers are a lot more stable.

[Fluffdaddy]

17. You Go Boyz...........This is way over my head.........But

It sounds like good news

[Tesibria]

19. I don't think he's averaging.

i think he's using rolling raw numbers, and I don't see how you can get that except somehow by extrapolating the number of people polled each day .. and then doing some sorta weird math of converting the percentages to numbers .... ?

but I really don't think he's averaging the three days

[Firespirit]

22. If Zogby uses the arithmetic mean..

..with minimal adjustments, then TIA's numbers are pretty darn close to the truth. I checked them using the simple arithmetic mean and came up with the same data for the day column.

The two shrub numbers in the daily column -- TIA's calculations -- that stand out to me are the 52.6 and 39.1. Whether these numbers are strictly correct or not, what's significant is that * experienced a pretty big one-day decline if the three-day average went down two whole points. Something outside the MoE.

Either that, or one of the other two values used in the average was off -- i.e., too high. TIA's value of 52.6 looked funny to me, since *'s numbers in both TIA's column and the Zogby data did not break 46% before that. That day's result, whatever it actually was, could be an outlier.

Either way, it's good news for us.

The 10-pt difference TIA has computed is also in line with the WaPo poll showing Kerry with a 10-point lead in some 13 swing states. I assume that is also an average; anyone ever find the details of that poll??

[demosincebirth]

25. Yea, thats the way I figured it.

Good news, you are always a purveyor of good news. Thanks TIA

[Wakeup_America]

26. You're right on the mark

Check out my numbers in posts 13 and 24. They match the Zogby's moving averages with lot less fluctuations.

[Skip Intro]

27. woops

[CatWoman]

23. You're so smart. THANKS!!!!!

:*

:hi:

[Skip Intro]

28. Way over my head, but

:thumbsup: :thumbsup: :thumbsup:

I knew * up by 8 (drudge/gallup/likely) was bullshit.

Thanx!!!!!!!!

[Fluffdaddy]

29. And Drudge could not wait tpo post it............I thought it was BS also