Warning

 

Close

Confirm Action

Are you sure you wish to do this?

Confirm Cancel
Member Login
Site Notices
11/22/2017 10:05:29 PM
Posted: 10/28/2004 12:03:18 PM EST
[Last Edit: 10/28/2004 12:13:38 PM EST by HiramRanger]
Link Posted: 10/28/2004 12:09:07 PM EST
Link Posted: 10/28/2004 12:09:36 PM EST
Link Posted: 10/28/2004 12:10:13 PM EST
Come ON, Tuesday! GET HERE already!
Link Posted: 10/28/2004 12:11:09 PM EST

Originally Posted By Zaphod:
Come ON, Tuesday! GET HERE already!



+1000
Link Posted: 10/28/2004 12:12:33 PM EST
According to Hannity, Bush is up 8 in FL, up 2 in PA and tied in NJ.

I like the trend.

Link Posted: 10/28/2004 12:13:56 PM EST

Originally Posted By garandman:
According to Hannity, Bush is up 8 in FL, up 2 in PA and tied in NJ.

I like the trend.




Like the technical traders say: The trend is your friend.

At least this one is.
Link Posted: 10/28/2004 12:15:21 PM EST
Link Posted: 10/28/2004 12:18:20 PM EST
That is damn good news. I get the feeling that all of these tracking polls are just bouncing around their margin of error. It is the trend that is important!
Link Posted: 10/28/2004 12:18:29 PM EST
For what it's worth you missed this one
www.time.com/time/election2004/article/0,18471,733715,00.html?cnn=yes

I dig the good news, but everyday this shit is teetering back and forth. I'm gonna be the next person to throw a brick, and I quit smoking 8 years ago.
Link Posted: 10/28/2004 12:21:18 PM EST
Not a bad turn-around for a tracking poll.


Now don't forget to vote.
Link Posted: 10/28/2004 12:24:17 PM EST
Link Posted: 10/28/2004 12:28:20 PM EST
Link Posted: 10/28/2004 12:29:31 PM EST
I hate to rain on everyone's parade, but from a mathematical standpoint, most of these poll numbers cannot tell us who is ahead. It would take a very, very long time to explain the math behind it, but if any of the polls are within the margin of error, the the poll cannot tell you anything. If it's 48% to 46% with margin of error of 3%, then you cannot say the person with 48% is ahead.

Besides that, many statisticians believe that modern phone polls are systemically flawed, even beyond the sample error.

With a race this close, only the election results themselves count. All others are bad estimates.
Link Posted: 10/28/2004 12:29:48 PM EST
At this point, I consider the National Polls to be almost irrelevant.
All I care about NOW, is seeing solid Bush trends in FLA, OH, MN, WI, CO, etc..
Link Posted: 10/28/2004 12:30:58 PM EST
[Last Edit: 10/28/2004 12:32:41 PM EST by PAEBR332]

Originally Posted By HiramRanger:

Originally Posted By slefelar:
That is damn good news. I get the feeling that all of these tracking polls are just bouncing around their margin of error. It is the trend that is important!



While most are within the MOE... let me state this with 100% conviction, if they were neck and neck and there was no true lead you would see the polls split between the candidates. The fact that they are all in our column except one tie is an indication that there is a definitive Bush lead and it is large enough that it is registering in all polls, albeit at a different level. You can not have this consistant of a trend if the lead is not there.



If you know your stats, you would see that HiramRanger is correct. Under the Central Limit Theorem, the mean of several sample means is likely to be the true mean. If this were truly tied, we should see about half the polls with Kerry ahead.
Link Posted: 10/28/2004 12:36:04 PM EST
Link Posted: 10/28/2004 12:37:05 PM EST

Originally Posted By PAEBR332:

Originally Posted By HiramRanger:

Originally Posted By slefelar:
That is damn good news. I get the feeling that all of these tracking polls are just bouncing around their margin of error. It is the trend that is important!



While most are within the MOE... let me state this with 100% conviction, if they were neck and neck and there was no true lead you would see the polls split between the candidates. The fact that they are all in our column except one tie is an indication that there is a definitive Bush lead and it is large enough that it is registering in all polls, albeit at a different level. You can not have this consistant of a trend if the lead is not there.



If you know your stats, you would see that HiramRanger is correct. Under the Central Limit Theorem, the mean of several sample means is likely to be the true mean. If this were truly tied, we should see about half the polls with Kerry ahead.



Correct...
Link Posted: 10/28/2004 12:40:32 PM EST
Here is another excellent read on polling in general and why Zogby is a hack:
jaycost.blogspot.com/
Link Posted: 10/28/2004 12:41:40 PM EST
The trend is your friend right up until it is not.

The trend is a fickle beast and it is always good to have backup or an escape plan.
Link Posted: 10/28/2004 12:42:26 PM EST
Hold on tight, most of the national polls gave Bush a lead in 2000.
Link Posted: 10/28/2004 12:43:26 PM EST
[Last Edit: 10/28/2004 12:44:44 PM EST by Dominus]
Be careful with the Central Limit Theorem. It requires each of the independent random variables to be identically distributed. Multiple polls from multiple sources at different times may not be. (meaning, different methodology).

If I took 10 polls today in the same city in the same method, then what you are saying is correct.

I should add the caveat that if you use more advanced and more generalized limiting theorems, you can make at least some conclusions. However, if someone has performed the analysis using these methods, and got useful results out of it, then I expect it would be headline news.

I don't have my stats books handy, so I don't want to get into a big arguement.
Link Posted: 10/28/2004 12:48:07 PM EST
I cannot WAIT to get this election done and over with so I can adorn my truck with bumper stickers mocking John Kerry's shady lifelong record - I'm active duty military and once Kerry completely leaves public office - I will make some comment on what a douchbag he is EVERY day.
Link Posted: 10/28/2004 12:50:57 PM EST
I believe you can Mock and bad mouth him all you want, until he's the president.
At least that's been my practice.
Link Posted: 10/28/2004 12:51:57 PM EST
The trend is our friend
Link Posted: 10/28/2004 12:54:09 PM EST
Oddly enough, I work with several active duty members who have no problem bitching about Bush now.....but they are your typical - my P$SSY hurts and I need to ETS now types. You know - the sniveling whining b$tches that CAN'T handle military life - but think they know what is best for the military. Fuck all the p#ssies and especially their mentor John F----ing Kerry - soon to be just Mr Teresa Heinz...........
Link Posted: 10/28/2004 12:54:23 PM EST
Get more Bush in 2005
Link Posted: 10/28/2004 1:05:07 PM EST
FWIW, El-Rushbo writes off Ohio because the R-Gov increased taxes too much. He says that we don't need it anyway.
Link Posted: 10/28/2004 1:08:28 PM EST

Originally Posted By Nozzleman:
FWIW, El-Rushbo writes off Ohio because the R-Gov increased taxes too much. He says that we don't need it anyway.



Rush predicated that with an "IF we lose Ohio..."

Top Top