DialIdol.com Forums
Home | Profile | Register | Active Topics | Members | Search | FAQ
Username:
Password:
Save Password Forgot your Password?

 All Forums
 DialIdol
 Predictions
 Given the margin of error, how is Taylor safe?
 New Topic  Topic Locked
 Printer Friendly

Author Previous Topic Topic Next Topic  

mgdpublic

28 Posts

Posted - 05/17/2006 :  11:53:34 AM  Show Profile
Taylor - moe < Katherine +moe.

supercal

34 Posts

Posted - 05/17/2006 :  12:03:39 PM  Show Profile
If you add Taylor's and Elliot's MOE (3.061) and add it to the lowest score (Elliot's 28.286), it comes up LESS than Taylor's score.

Elliot - 28.286 + 3.061 = 31.347

Taylor - 31.638

31.347 < 31.368

Taylor is not within the statistical MOE, therefore he is considered SAFE according to Dialidol guidelines.

Go to Top of Page

foobar

Suriname
67 Posts

Posted - 05/17/2006 :  2:40:04 PM  Show Profile
quote:
Originally posted by supercal

If you add Taylor's and Elliot's MOE (3.061) and add it to the lowest score (Elliot's 28.286), it comes up LESS than Taylor's score.

Elliot - 28.286 + 3.061 = 31.347

Taylor - 31.638

31.347 < 31.368

Taylor is not within the statistical MOE, therefore he is considered SAFE according to Dialidol guidelines.





The margins of error are not calculated correctly. They should actually be about 0.08.

There may still be systematic sources of error.

Here's how to calculate it....

Fundamentally, you use the formula

MOE = 2.58 * Standard Error

For this type of sampling process, Standard error is
SQRT ( p * (1-p) / N)
N is the number of Dialidol attempts for a specific contestant
P is the busy signal percentage for that contestant.

So, with 80% busy and 220000 Dialidol votes (ballpark for Taylor)
sqrt( 0.8 * 0.2 / 220000) = 0.08%
Which makes the standard error 0.2%

However, the Dialidol scores are about 37% of the Busy percentages, so you need to scale it down to 0.08 relative to the Dialidol scores (Jim coulda corrected for this too, but the scores work regardless of whether they are scaled down or not).

If the statistical model is 100% valid and the text votes have the same statistical distribution as the phone votes, Elliott is gone, for certain, tonite. Elliot is 0.473 behind Kath, and the MOE is 0.08. I think he had poor song choice and she didn't forget the words and look like an idiot this week.

Lot of caveats there, though...but I have faith in Dialidol at this point.

We shall see....



Uh Huh
Go to Top of Page
  Previous Topic Topic Next Topic  
 New Topic  Topic Locked
 Printer Friendly
Jump To:
DialIdol.com Forums © www.DialIdol.com Go To Top Of Page
Snitz Forums 2000
Google
DialIdol.com is NOT affiliated in any way with FOX TV ®, AMERICAN IDOL ®, or FremantleMedia North America, Inc.