SI
SI
discoversearch

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor.  We ask that you disable ad blocking while on Silicon Investor in the best interests of our community.  If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.
Politics : Foreign Affairs Discussion Group

 Public ReplyPrvt ReplyMark as Last ReadFilePrevious 10Next 10PreviousNext  
To: Neocon who wrote (150190)11/1/2004 11:08:59 AM
From: Jim Willie CB  Read Replies (3) of 281500
 
wrong on all counts, Neocon
hear this, from a PhD in Statistics

1. a margin of error pertains to a 95% confidence interval on the individual preference stated... the spread difference must be multiplied by the square root of 2 (which is 1.414--, or 41.4% wider, not twice as wide)

2. my point exactly, which you missed... most polls should NOT exceed the +/-3%, as in 95% of them should be within that margin... routinely one poll out of only a handful exceed, far more than 1 in 20 polls... a lot of polls should NOT exceed the margin of error... of course it is a function of sample size, the larger its size, the narrower the margin of error, all factored in... so nonsensical statement on your part

3. mere statistical anomalies are the norm, the same point you missed entirely... you should have one anomaly every 15 to 25 polls, not every four or five... anomalies have become the norm in prez polling, which means you fail to comprehend the statistical nature of sampling and inference entirely

4. reweighting employs stratification methods, which introduces tremendously powerful statistical accuracy potential... what you are saying is that they make them wider apart... a statistical theorem can be shown that stratified samples produce markedly lower margins of error, and can be done so in an optimal fashion... of course "likely voters" approaches are key to my entire point, wherein pollsters use their own method to bring about the result they want, like Gallup being the top Bush poll almost every single time (it should be top of 5 polls only once every 5 polling exercises)... likely voter interpretation and execution is central to the arbitrary nature of the polling, and ripe with bias, which is my point

check out an article written on polling complexity
"Primer on Presidential Polls"
financialsense.com

I hope you found this educational
by the way, in my view "neocon" is a synonym for "fascist"
if you are a true Neocon, you will ignore all facts and maintain your wrong views, by fiat and by force if necessary

as always, I hope Republicans fail to find their way to the voting centers tomorrow, for whatever reason
- an extra few bucks to make, or to swindle, or to steal
- too busy intimidating others to exercise their own rights
- tied up making millions in war material businesses
- distracted by their legal challenges of voter registrations

my statistical views are mathematical, not political

/ jim

p.s. how do you like the newest Republican method ???
challenge of voter registration based upon weight of paper
happening in Ohio now
much talk about freedom, but challenges of key rights which freedom gaurantees, like voting
incredibly hypocritical

calling black voters by telephone from Voter Rights Group to vote on Wednesday
dont tell me it aint happening, I have a black friend who has second hand knowledge of it happening
Report TOU ViolationShare This Post
 Public ReplyPrvt ReplyMark as Last ReadFilePrevious 10Next 10PreviousNext