Rand is polling > 15% with the under 50 age group

[limequat]

Of course, they won't show directly in any poll, but you can calculate it for yourself based on the latest CNN Poll

CNN marks confidence levels of >8.5% as "NA". For a group to be marked NA with a sample size of 465, the NA sample has to be less than 133.
The total republican sample size is 465 of which Rand get's 5% of the vote.
He gets 1% with the "over 50" age group.

We can take those numbers to write this equation...

Samples Under 50 * X + Samples Over 50 * Percentage = Total Samples * Percentage
133 * X + 317 * 1% = 465 * 5%

X = 15%

 

[CaptUSA]

+rep for this.

And let's not forget, this is only a snap shot. 2/3rds of voters are still undecided and only pick a candidate when they say, "Well, if you had to pick a candidate today..."

 

[Todd]

Math imbecile here;

Can you explain it to us? if the unknown is X where do we get the 317

 

[limequat]

Math imbecile here;

Can you explain it to us? if the unknown is X where do we get the 317

I glossed over that. CNN lists the confidence level for the over 50 group as 5.5%. This correlates to a sample of 317.

 

[whoisjohngalt]

I glossed over that. CNN lists the confidence level for the over 50 group as 5.5%. This correlates to a sample of 317.

Lol, you still kind of glossed over it. But I really appreciate you doing this.

So 19 million people voted in the last GOP primary. That's our Population Size, the population they are trying to sample.

I could get into the actual math, but the formula is this: z* sqrt(p(1-p)/n) where p is the population size and n is the sample size.

Much easier, use this calculator: http://americanresearchgroup.com/moe.html

Plug in your 19,000 pop and start guessing at the sample size until your margin of error hits 5.5. Hint: it's 317

 

[limequat]

Lol, you still kind of glossed over it. But I really appreciate you doing this.

So 19 million people voted in the last GOP primary. That's our Population Size, the population they are trying to sample.

I could get into the actual math, but the formula is this: z* sqrt(p(1-p)/n) where p is the population size and n is the sample size.

Much easier, use this calculator: http://americanresearchgroup.com/moe.html

Plug in your 19,000 pop and start guessing at the sample size until your margin of error hits 5.5. Hint: it's 317

Thanks for elaborating!

Also my understanding is that as the population gets larger, it becomes less relevant. So with a population of millions, we are really only looking at sample size.

 

[JohnGalt23g]

Uhhh, I was told there wouldn't be any math on the thread...

 

[whoisjohngalt]

Thanks for elaborating!

Also my understanding is that as the population gets larger, it becomes less relevant. So with a population of millions, we are really only looking at sample size.

Certainly in electoral politics. Technically I would say that as the ratio of sample to population shrinks, the significance of the population size decreases.

 

[specsaregood]

It is worth noting why they don't show the % for groups with such a small sample size; it is because the value is unreliable.

 

[Jonderdonk]

Uhhh, I was told there wouldn't be any math on the thread...

I think you're looking for the Bernie Sanders forums.

 

[farrar]

Lol, you still kind of glossed over it. But I really appreciate you doing this.

So 19 million people voted in the last GOP primary. That's our Population Size, the population they are trying to sample.

I could get into the actual math, but the formula is this: z* sqrt(p(1-p)/n) where p is the population size and n is the sample size.

Much easier, use this calculator: http://americanresearchgroup.com/moe.html

Plug in your 19,000 pop and start guessing at the sample size until your margin of error hits 5.5. Hint: it's 317

weird... I just finsihed learning about this in statistics less than an hour ago... then i come on here and i get a recap. lol.

 

[Chester Copperpot]

glad you did the math on this.. it would make my brain hurt. lol

 

[eleganz]

Feeling thankful for the liberty math nerd squad. <3

 

[whoisjohngalt]

Feeling thankful for the liberty math nerd squad. <3

I'm a statistical analyst professionally and very much enjoy these types of things, but the real beast of RPF when it comes to all things statistical is malkusm. He is a statistical genius. I'd like to have him start digging into some of these polls.

 

[r3volution 3.0]

Very encouraging news

"You must spread some Reputation around before giving it to limequat again"

 

[Todd]

I'm a statistical analyst professionally and very much enjoy these types of things, but the real beast of RPF when it comes to all things statistical is malkusm. He is a statistical genius. I'd like to have him start digging into some of these polls.

Haven't seen him here for some time. Plus one on wanting to get his take.

 

[opinionatedfool]

Great post OP

 

[KingNothing]

I'm a statistical analyst professionally and very much enjoy these types of things, but the real beast of RPF when it comes to all things statistical is malkusm. He is a statistical genius. I'd like to have him start digging into some of these polls.

Speaking of statistical geniuses, remember the guy who absolutely was not one who repeatedly claimed that Ron only lost because of "algorithmic vote flipping," and that he stumbled across undeniable proof that this was the case?

That was really, really weird.

 

[EBounding]

Of course, they won't show directly in any poll, but you can calculate it for yourself based on the latest CNN Poll

CNN marks confidence levels of >8.5% as "NA". For a group to be marked NA with a sample size of 465, the NA sample has to be less than 133.
The total republican sample size is 465 of which Rand get's 5% of the vote.
He gets 1% with the "over 50" age group.

We can take those numbers to write this equation...

Samples Under 50 * X + Samples Over 50 * Percentage = Total Samples * Percentage
133 * X + 317 * 1% = 465 * 5%

X = 15%

Now do it the Common Core way.

 

[devil21]

I've been saying that the polls being released are directly targeted to and comprised of old people that live in front of Fox News all day and night. It's fair to say that the old people are almost guaranteed to vote, however these polls are not even close to a broad representation of all voters. This is a large part of the psyop the media runs to create the image of who is "winning". Targeting polling to people they KNOW will respond like pavlov's dogs to Fox propaganda.

No one can convince me that the first two debates, with 20+ million viewers each, have resulted in Rand only receiving 2-3% across the board. No way.