A real improvement or a random noise?
       |           

Welcome, Guest. Please login or register.
Did you miss your activation email?
March 29, 2024, 10:27:27 AM
News: Election Simulator 2.0 Released. Senate/Gubernatorial maps, proportional electoral votes, and more - Read more

  Talk Elections
  Election Archive
  Election Archive
  2004 U.S. Presidential Election
  A real improvement or a random noise?
« previous next »
Pages: [1] 2
Author Topic: A real improvement or a random noise?  (Read 3012 times)
Shira
Jr. Member
***
Posts: 1,858


Show only this user's posts in this thread
« on: September 24, 2004, 11:57:14 AM »



According to Rasmussen daily sampling, Bush has 47.4% and Kerry 46.5%.
Is it a real improvement for Kerry or a random statistical noise?
I believe that Kerry improved a little bit in the last few days, but the nationwide gap is probably around the 3%.
Logged
Sam Spade
SamSpade
Atlas Star
*****
Posts: 27,547


Show only this user's posts in this thread
« Reply #1 on: September 24, 2004, 12:03:50 PM »

Feels like a blip to me.  Of course, we will know more in about 3 days when the sample rolls out of the mix.

The best way to look at Rasmussen is to compare his weekly averages, IMO and try and spot trends.
Logged
Fritz
JLD
Junior Chimp
*****
Posts: 5,668
United States


Show only this user's posts in this thread
« Reply #2 on: September 24, 2004, 12:12:01 PM »

Today- Bush +0.9
Sept 23- Bush +3.6
Sept 22- Bush +4.0
Sept 21- Bush +2.8
Sept 20- Bush +2.8
Sept 19- Bush +1.7
Sept 18- Bush +4.6

I'd like to see <1% for three days before I say a trend has developed.
Logged
Fmr. Gov. NickG
NickG
Junior Chimp
*****
Posts: 8,150


Political Matrix
E: -8.00, S: -3.49

Show only this user's posts in this thread
« Reply #3 on: September 24, 2004, 12:37:08 PM »


I wouldn't but any faith whatsoever in a one day blip in Rasmussen polling.  In fact, I'm not sure I trust the poll at all.  Rasmussen only had Bush 1 a week after the RNC when most polls had him up 7-12%, but had him up 4 for most of this week, when most polls showed him up losing his margin.  The Rasmussen track has been weirdly stable all year....It hasn't really shown any sustained trends at all of more than a point or two.
Logged
MODU
Atlas Star
*****
Posts: 22,024
United States


Show only this user's posts in this thread
« Reply #4 on: September 24, 2004, 12:58:15 PM »


That's why I don't look at his actual figures, but rather the trends.  Even with bad polling, if the polling method remains the same, it should pick up trends.  If the gap keeps increasing in Bush's favor, then Bush is doing well.  If the gap decreases in Kerry's favor, then Bush is losing ground.  Just that simple.  Smiley
Logged
The Vorlon
Vorlon
YaBB God
*****
Posts: 4,657


Political Matrix
E: 8.00, S: -4.21

Show only this user's posts in this thread
« Reply #5 on: September 24, 2004, 01:12:47 PM »



According to Rasmussen daily sampling, Bush has 47.4% and Kerry 46.5%.
Is it a real improvement for Kerry or a random statistical noise?
I believe that Kerry improved a little bit in the last few days, but the nationwide gap is probably around the 3%.


If its a blip it will stay with the 3 day rolling average for... 3 days.

If on Monday Bush bounces back to +4 (when the Thursday night sample rolls off) we know it was a blip, if not.. then not Smiley

I'll let you know on MOnday.

For what it is worth, a very strong Bush sample rolled OFF today, the last 3 daily samples have all been within about 1.5% of tied - but they have also shown a fairly large blip towards the dems in party ID.

Hard to say what, if anyting, it means.
Logged
Shira
Jr. Member
***
Posts: 1,858


Show only this user's posts in this thread
« Reply #6 on: September 24, 2004, 05:27:15 PM »
« Edited: September 24, 2004, 05:41:13 PM by Shira »



According to Rasmussen daily sampling, Bush has 47.4% and Kerry 46.5%.
Is it a real improvement for Kerry or a random statistical noise?
I believe that Kerry improved a little bit in the last few days, but the nationwide gap is probably around the 3%.


If its a blip it will stay with the 3 day rolling average for... 3 days.

If on Monday Bush bounces back to +4 (when the Thursday night sample rolls off) we know it was a blip, if not.. then not Smiley

I'll let you know on MOnday.

For what it is worth, a very strong Bush sample rolled OFF today, the last 3 daily samples have all been within about 1.5% of tied - but they have also shown a fairly large blip towards the dems in party ID.

Hard to say what, if anyting, it means.

If we compare 3 polls of Time we'll see that on 9/2 Bush had 10% margin. On 9/9 the margin was 12% while yesterday (9/23) it was 4%.
Further more: Does Bush deserve a second term or someone else?
On 9/9 it was 52:45 in favor of Bush, while yesterday it was 49:47 in favor of someone else.
Is it all random or is it the beginnig of some shift?

And another sign:
On 9/6 the EV as per Rasmussen were 213:175 in favor of Bush.
On 9/21 these numbers were 213:211

I can not argue about the professionalism of Rasmussen but the methodology in these two dates was the same



Logged
Sam Spade
SamSpade
Atlas Star
*****
Posts: 27,547


Show only this user's posts in this thread
« Reply #7 on: September 24, 2004, 05:42:47 PM »

Maybe.  It's also quite possible that the last Time poll was an outlier, (which was my gut feeling at the time)because no one else showed the same results they did, except for Gallup LV and Gallup LV is a very different poll and typically isn't accurate till near the election.

I also make my judgment based on the fact that no other polls out in the past week (AP-Ipsos, CBS, IBD, Fox, Zogby) seem to be showing this movement.

Time will tell, however.  BTW, the margin in the latest Time poll is 6 point for Bush, not 4
Logged
Shira
Jr. Member
***
Posts: 1,858


Show only this user's posts in this thread
« Reply #8 on: September 24, 2004, 05:48:21 PM »
« Edited: September 24, 2004, 05:50:26 PM by Shira »


Time will tell, however.  BTW, the margin in the latest Time poll is 6 point for Bush, not 4

Go to
http://www.pollingreport.com/wh04gen.htm


The 6% margin is with Nader who virtually gets 5%
Logged
Reds4
Jr. Member
***
Posts: 789


Show only this user's posts in this thread
« Reply #9 on: September 24, 2004, 07:12:42 PM »

In my opinion Kerry is coming back some, which is to be expected. Unlike many of my GOP friends I think this race is going to be dead even again before November 2nd. The debates and events between now and the election should decide it.
Logged
ToomeyforPres
Rookie
**
Posts: 17


Show only this user's posts in this thread
« Reply #10 on: September 24, 2004, 07:14:11 PM »

Bush is up at least ten, and will probably win by more on election day
Logged
Sam Spade
SamSpade
Atlas Star
*****
Posts: 27,547


Show only this user's posts in this thread
« Reply #11 on: September 24, 2004, 07:17:57 PM »

It's 4 among RV whether or not Nader is included, which is what Polling Report is reporting.  It's 6 among LV whether or not Nader is included, which is what Time is reporting.
Logged
The Vorlon
Vorlon
YaBB God
*****
Posts: 4,657


Political Matrix
E: 8.00, S: -4.21

Show only this user's posts in this thread
« Reply #12 on: September 25, 2004, 01:57:49 PM »
« Edited: September 25, 2004, 02:00:15 PM by The Vorlon »

A breakout of the Daily Bot track - day by day samples:

Logged
agcatter
agcat
YaBB God
*****
Posts: 3,740


Show only this user's posts in this thread
« Reply #13 on: September 25, 2004, 02:15:37 PM »

Are these day to day numbers all already weighted 39-35- 26?
Logged
Shira
Jr. Member
***
Posts: 1,858


Show only this user's posts in this thread
« Reply #14 on: September 25, 2004, 02:22:54 PM »
« Edited: September 25, 2004, 02:27:47 PM by Shira »

A breakout of the Daily Bot track - day by day samples:



As oppose to what your graph shows, never in September was Kerry ahead of Bush. That's what in your graph on 9/14, 9/19, 9/24.
Further more, the numbers of 9/25 (today) are 47.9:46.3 and not 49.8:45.4
Logged
Sam Spade
SamSpade
Atlas Star
*****
Posts: 27,547


Show only this user's posts in this thread
« Reply #15 on: September 25, 2004, 02:25:17 PM »

So, right now for the week of Sept. 19-25, we have Bush 48.04 to Kerry 45.83 or a lead of roughly Bush +2.27 for the week.

The previous week of Sept. 12-18, it was Bush 48.44 to Kerry 45.31, or a lead of roughly Bush +3.13 for the week.

I'll buy that from a poll weighted 39D-35R-26I as such.  There might be some movement here, but it's sort of hard to tell whether it's that or just variance under the MOE, which is still could easily be, even with 7,000 LV samples.
Logged
agcatter
agcat
YaBB God
*****
Posts: 3,740


Show only this user's posts in this thread
« Reply #16 on: September 25, 2004, 02:29:19 PM »
« Edited: September 25, 2004, 02:30:25 PM by agcat »

I understood that the numbers in Vorlon's chart were the numbers from each individual night's polling sample.

Nobody else knows how to figure the DAILY results.  

Shira, I think you are referring to the three day rolling number.
Logged
Sam Spade
SamSpade
Atlas Star
*****
Posts: 27,547


Show only this user's posts in this thread
« Reply #17 on: September 25, 2004, 02:30:54 PM »

That was his sample for the specific day, (9/25) where it is 49.8 to 45.4.

The average of the last three days (9/23-25) is 47.9 to 46.3.

As to Kerry leads on certain days, if we were to theoretically say that Bush's lead under this weighting is roughly 2.5% (which is a reasonable guess), then each daily sample would roughly have a MOE of +/- 3%, which means that we would get a Kerry lead somewhere on the plotted graph every once in a while (not doing the math to figure it out specifically).
Logged
Shira
Jr. Member
***
Posts: 1,858


Show only this user's posts in this thread
« Reply #18 on: September 25, 2004, 02:38:36 PM »

That was his sample for the specific day, (9/25) where it is 49.8 to 45.4.

The average of the last three days (9/23-25) is 47.9 to 46.3.

As to Kerry leads on certain days, if we were to theoretically say that Bush's lead under this weighting is roughly 2.5% (which is a reasonable guess), then each daily sample would roughly have a MOE of +/- 3%, which means that we would get a Kerry lead somewhere on the plotted graph every once in a while (not doing the math to figure it out specifically).

But if you look into the Rasmussen stuff (  http://www.rasmussenreports.com/Presidential_Tracking_Poll.htm ) you'll see that on 9/7 they were tied 47.3% and in all other days Bush was ahead.
Does Rasmussen show average of the last three days and The Varlion derives the single date from it?
Logged
agcatter
agcat
YaBB God
*****
Posts: 3,740


Show only this user's posts in this thread
« Reply #19 on: September 25, 2004, 02:50:40 PM »

Somehow, some way, he is able to figure the one day samples.  No one on this board knows how he figures it out, but I'd bet money his numbers are dead on.
Logged
Shira
Jr. Member
***
Posts: 1,858


Show only this user's posts in this thread
« Reply #20 on: September 25, 2004, 03:14:31 PM »
« Edited: September 25, 2004, 03:26:12 PM by Shira »

Somehow, some way, he is able to figure the one day samples.  No one on this board knows how he figures it out, but I'd bet money his numbers are dead on.

OK
Let X22, X23, X24 and X25 be the margins on 09/22, 09/23, 09/24 and 09/25.

So we have (X25 + X24 + X23)/3 = 1.6
And (X24 + X23 + X22)/3 = 0.9

or

X25 + X23 + X24 = 4.8
X22 + X23 + X24 = 2.7

By subtracting we’ll get X25 – X22  = 2.1
which means that the margin on 9/25 was greater than the margin on 9/22 by 2.1% and that’s what his chart shows. In order to progressively calculate the absolute one day number he has to know the absolute margin number of one day or may be two consecutive days ( I have to think about it)

It's even simpler. The assumption is that if a 3 days margin avarage shifted 0.7 EACH day between 9/22 and 9/25 so the total shift is 2.1 and that's what his chart really shows.
Logged
The Vorlon
Vorlon
YaBB God
*****
Posts: 4,657


Political Matrix
E: 8.00, S: -4.21

Show only this user's posts in this thread
« Reply #21 on: September 25, 2004, 03:41:54 PM »

Somehow, some way, he is able to figure the one day samples.  No one on this board knows how he figures it out, but I'd bet money his numbers are dead on.

+/- 0.3% actually (Rounding error) but close enough Smiley

Actually even at 1000 interviews per night, the amount of random "noise" is fairly high.

Each individual day will have a MOE of 3.1%, however since each candidate is subject to error, the actual error on the reported lead is 1.414 * 3.1 => +/- 4.4% on the "lead"  In a tracking poll you also have two error inputs, because the random error of the sample dropping off is also a factor.

For example, if pro-Kerry "random error" drops off the 3 day sample, the same day some pro-Bush random error rolls one, you a have two blips operating together.  Sometimes the errors cancel, but sometimes they add.

On average the daily difference in error (between the sample rolling on and off) will be +/- 6.2% 19/20 times.

At the 50% confidence level, this is about +/- 2.1%

Bottom line:

If the 3 day average moves 0.7% there is a 50/50 chance it is just pure random noise.
A 1% shift has a 1/3rd chance of being pure noise.

Once or twice a month, we will get a 1 day shift over 2% that is just nothing but random noise.

If you look at anything in Rasmussen, look at the 7 day average.


Logged
The Vorlon
Vorlon
YaBB God
*****
Posts: 4,657


Political Matrix
E: 8.00, S: -4.21

Show only this user's posts in this thread
« Reply #22 on: September 25, 2004, 03:43:44 PM »

Are these day to day numbers all already weighted 39-35- 26?

yes, the single day vaues are weighted to 39/35/26.  I can show you the unweighted values as well, but the bounce around a ton and are not particularly useful.

Rasmussen actually has a rather large "Day of the week" issue.  He weights a ton of it away with his 39/35/26, but in the raw data it is rather a meaningful effect.
Logged
Shira
Jr. Member
***
Posts: 1,858


Show only this user's posts in this thread
« Reply #23 on: September 25, 2004, 03:47:14 PM »



According to Rasmussen daily sampling, Bush has 47.4% and Kerry 46.5%.
Is it a real improvement for Kerry or a random statistical noise?
I believe that Kerry improved a little bit in the last few days, but the nationwide gap is probably around the 3%.


If its a blip it will stay with the 3 day rolling average for... 3 days.

If on Monday Bush bounces back to +4 (when the Thursday night sample rolls off) we know it was a blip, if not.. then not Smiley

I'll let you know on MOnday.

For what it is worth, a very strong Bush sample rolled OFF today, the last 3 daily samples have all been within about 1.5% of tied - but they have also shown a fairly large blip towards the dems in party ID.

Hard to say what, if anyting, it means.

Correct me if I am wrong.
You have to know the absolute numbers of arbitrarry two consecutive dates in order to run this process of "deavaraging"
Logged
The Vorlon
Vorlon
YaBB God
*****
Posts: 4,657


Political Matrix
E: 8.00, S: -4.21

Show only this user's posts in this thread
« Reply #24 on: September 25, 2004, 03:49:30 PM »
« Edited: September 25, 2004, 03:51:03 PM by The Vorlon »



According to Rasmussen daily sampling, Bush has 47.4% and Kerry 46.5%.
Is it a real improvement for Kerry or a random statistical noise?
I believe that Kerry improved a little bit in the last few days, but the nationwide gap is probably around the 3%.


If its a blip it will stay with the 3 day rolling average for... 3 days.

If on Monday Bush bounces back to +4 (when the Thursday night sample rolls off) we know it was a blip, if not.. then not Smiley

I'll let you know on MOnday.

For what it is worth, a very strong Bush sample rolled OFF today, the last 3 daily samples have all been within about 1.5% of tied - but they have also shown a fairly large blip towards the dems in party ID.

Hard to say what, if anyting, it means.

Correct me if I am wrong.
You have to know the absolute numbers of arbitrarry two consecutive dates in order to run this process of "deavaraging"

No Smiley

Rounding error accumulates if you do that.  It 46.7 really 46.66 or 46.74..?
Logged
Pages: [1] 2  
« previous next »
Jump to:  


Login with username, password and session length

Terms of Service - DMCA Agent and Policy - Privacy Policy and Cookies

Powered by SMF 1.1.21 | SMF © 2015, Simple Machines

Page created in 0.044 seconds with 12 queries.