Battleborne
Terminator
What can I say...
other than I always believe there is room for improvement on any subject!
other than I always believe there is room for improvement on any subject!
40% of Americans Want Muslims to have special ID
- Thread starter BRUTALITOPS
- Start date
T
TRGLDTE
Guest
That depends fully on the confidence level and margin of error the polltaker seeks.
T
TRGLDTE
Guest
Even a sample of 100% of the population has an associated error term- the error term does not go to zero. (A physicist would recognize Heisenberg hiding in the weeds.)You must be the one under the influence...I used the term 300m(as in million)
Grind said only 300 were needed for a accurate poll as in 300 not 300m
and the 300m was in reference to the US population and noted 10% or 30m would in my opinion be needed for a accurate marker...providing the questions asked were not biased! as in statistics the more the merrier for accurate results!
So for a sample size of 1000 out of ~300 million the m.o.e. is roughly +/-3.2% at a 95% confidence level.
maineman
Banned
the fact remains, as I said, this statement of yours is inaccurate. The poll has a sufficient sample size to be totally valid and you obviously did not take stats 101 or you would know this.
Battleborne
Terminator
Thats what they say...
however until and unless someone actually does a poll of say my 10% theory some 30m...this point is a moot one!
however until and unless someone actually does a poll of say my 10% theory some 30m...this point is a moot one!
Battleborne
Terminator
Wrong again...
I did take stats 101...this is my hypothesis...like it or not...prove me wrong do a 10% poll of 30m...so far no one has!
I did take stats 101...this is my hypothesis...like it or not...prove me wrong do a 10% poll of 30m...so far no one has!
T
TRGLDTE
Guest
Quote:
Originally Posted by Troglodyte27
Even a sample of 100% of the population has an associated error term- the error term does not go to zero. (A physicist would recognize Heisenberg hiding in the weeds.)
So for a sample size of 1000 out of ~300 million the m.o.e. is roughly +/-3.2% at a 95% confidence level.
Absolutely incorrect.
Originally Posted by Troglodyte27
Even a sample of 100% of the population has an associated error term- the error term does not go to zero. (A physicist would recognize Heisenberg hiding in the weeds.)
So for a sample size of 1000 out of ~300 million the m.o.e. is roughly +/-3.2% at a 95% confidence level.
however until and unless someone actually does a poll of say my 10% theory some 30m...this point is a moot one!
Absolutely incorrect.
Battleborne
Terminator
Why....???
You are basing your theory on 'probability'...all this means is it is probable...not a proven fact...and by the way this only works on inanimate objects such as a two sided coin...even then it works out to fifty fifty...when working with the humane mind such as in polls one needs as close to one-hundred percent as one can feasably get in order to compensate for the individual mind! This is my hypothesis...whether or not you buy it...do the 10% poll then get back to me!
You are basing your theory on 'probability'...all this means is it is probable...not a proven fact...and by the way this only works on inanimate objects such as a two sided coin...even then it works out to fifty fifty...when working with the humane mind such as in polls one needs as close to one-hundred percent as one can feasably get in order to compensate for the individual mind! This is my hypothesis...whether or not you buy it...do the 10% poll then get back to me!
T
TRGLDTE
Guest
Where did I claim it a proven fact? In fact I recognized and acknowledged the limits of said measure, too bad you are not capable of understanding that. And if you don't see that in what I have written already, then you haven't the sense to understand that a pollster doesn't need to do a 10% study. As you pointed out, all it does is waste money. (Nice strawman / deflect from the provable fact that you know jack shit about stats though). Go for it - do your 10% poll and prove that everyone else is wrong as you contend - I'm not your lackey, do your own work.You are basing your theory on 'probability'...all this means is it is probable...not a proven fact...and by the way this only works on inanimate objects such as a two sided coin...even then it works out to fifty fifty...when working with the humane mind such as in polls one needs as close to one-hundred percent as one can feasably get in order to compensate for the individual mind! This is my hypothesis...whether or not you buy it...do the 10% poll then get back to me!
As to animate / inanimate, human(e) mind, sorry but you're wrong again. The behavior of humans en masse can be extrapolated quite well from groups smaller than the whole. The very reason to use groups is to attenuate the "individual mind" outlier data points.
Next?
Last edited:
OrnotBitwise
Watermelon
I've noticed this tendency among conservatives. Any time a poll indicates a result they don't like they suddenly don't believe that probability is relevant.
Battleborne
Terminator
Did I....
Call you a name for expressing your view?...NO!...so why are you calling me a name for putting forth my hypothesis? As a matter of fact polls are done in the small range to save money and put forth the pollers agenda..."They Can't handle the truth" And I never said a larger number was a fact...just that it would give a more accurate marker! And this would cost the pollers way more thus they would stop with all the useless polls to begin with!
Call you a name for expressing your view?...NO!...so why are you calling me a name for putting forth my hypothesis? As a matter of fact polls are done in the small range to save money and put forth the pollers agenda..."They Can't handle the truth" And I never said a larger number was a fact...just that it would give a more accurate marker! And this would cost the pollers way more thus they would stop with all the useless polls to begin with!
Battleborne
Terminator
This applies...
to both sides of the aisle...I on the other hand am a Independent thinker...I do not like polls period!They are useless!
to both sides of the aisle...I on the other hand am a Independent thinker...I do not like polls period!They are useless!
maineman
Banned
Call you a name for expressing your view?...NO!...so why are you calling me a name for putting forth my hypothesis? As a matter of fact polls are done in the small range to save money and put forth the pollers agenda..."They Can't handle the truth" And I never said a larger number was a fact...just that it would give a more accurate marker! And this would cost the pollers way more thus they would stop with all the useless polls to begin with!
actually what you SAID was:
"How can a poll be accurate when a sample of just 1000 out of some 300m is used as a marker...statistics 101 would put a damper on any poll taken with this marker...you would need at least 10% of the population to make a poll anywhere close to accurate...say 30m or so...not gonna happen so [polls are for the most part useless]"
When the fact is: a poll of 1000 folks can be quite accurate. No reputable statistician would suggest that you need 10% of total population as a sample size to "make a poll anywhere close to accurate" You were wrong.... again.
It seems as if it is habit forming for you.
Annie
Not So Junior Member
I'd say already formed.
bb, you are a moron. You don't need 10% to be accurate. Sample sizes do not have to be large as say 10%.
Annie
Not So Junior Member
I think we have a consensus here, I'm 'conservative' (though by another's definition I may be liberal, so who knows?), and agree.
Maineman: the fact is: a poll of 1000 folks can be quite accurate. No reputable statistician would suggest that you need 10% of total population as a sample size to "make a poll anywhere close to accurate" You were wrong.... again. It seems as if it is habit forming for you.
Grind: bb, you are a moron. You don't need 10% to be accurate. Sample sizes do not have to be large as say 10%.
Runyon: I think we have a consensus here, I'm 'conservative' (though by another's definition I may be liberal, so who knows?), and agree.
It seems we have a consensus that Battleboob is the board's resident moron.
Grind: bb, you are a moron. You don't need 10% to be accurate. Sample sizes do not have to be large as say 10%.
Runyon: I think we have a consensus here, I'm 'conservative' (though by another's definition I may be liberal, so who knows?), and agree.
It seems we have a consensus that Battleboob is the board's resident moron.
Battleborne
Terminator
Oh my...
I have been chastized by a panel of self proclaimed experts...
I have been chastized by a panel of self proclaimed experts...
maineman
Banned
you have been chastised by people who understand polling and statistics... topics of which you are obviously ignorant. One does not need to be an "expert" to know that your claim that a sample size of 10% of the population is required for polling data to have validity is silly. Like YOU said, it is covered in Statistics 101. If you took the class, you either flunked it or paid someone to take the tests for you... because you are clearly clueless on the subject.