Gun Rights

Everyone should have an RPG, you may never know if a tank might drive up to you and mug you.

1000 randomly sampled people given a country of 315 million IS approximately a 3% margin of error
Actually it is a very large margin of error. Almost like a 100% margin of error.

Blackflag: Have you ever taken an entry-level statistics paper? Seriously? Or are you disputing the statistical formula for a margin of error?

Actually I kinda figure I'd better show the working for anyone who hasn't done stats before.

Margins of error are generally derived from something called a confidence interval. The maths is a bit complex but for 99.99% of the surveys etc out there, the formula can be reduced to:
z* X sqrt( p* X ( 1 - p* ) / n )
(the main case where this does not hold true is if the proportion of the sample is less than 10 people or so)

z* is what's called the z-score, and it depends on your confidence interval. For 95% confidence, it is always 1.9.
p* is your proportion. So, if you want to see the margin of error on your claim that 80% of people are cool, then your p* is 0.8.
n is your sample size. In this case, 1000.

In this particular case:
Margin of Error (95% confidence)
z* X sqrt( p* X ( 1 - p* ) / n )
1.9 X sqrt( 0.49 X ( 1 - 0.49 ) / 1000 )
1.9 X sqrt( 0.49 X 0.51 / 1000 )
1.9 X sqrt( 0.2499 / 1000 )
1.9 X sqrt( .0002499 )
1.9 X 0.0158
= 0.03

admin: Also, this requires that the sample be TRULY random, which frankly almost no samples are. By I wouldn't fault them on that without specific evidence.

admin: Actually I kinda figure I'd better show the working for anyone who hasn't done stats before.
That is some really pure meth you are smoking

I don't care what you think you know, if you sample 1000 random people out of a pool of 315 million, you are going to get wildly inaccurate results 100% of the time. That is reality, not your version of it.

Or are you disputing the statistical formula for a margin of error?
Yes, I am disputing the version you provided.

It is extremely illogical. to think that 1000 randomly sampled people out of 315 million people are going to provide an accurate statistic for the total population 97% of the time.

I know that they actually sampled this with over 200 million people in a census, and there was a 29% difference from the "truthaboutguns.org" survey.

Blackflag: Would you like to debate that accurate statistical inference exists outside of my reality?

Also, are you sure you're not a 13 year old with who's never heard of a t-test in their life?

Blackflag: 1000 randomly sampled people out of 315 million people are going to provide an accurate statistic for the total population 97% of the time
Also, this demonstrates that you don't know the statistical meaning of the words "margin of error"

admin: Would you like to debate that accurate statistical inference exists outside of my reality?
No, because there is no point in debating the blatantly obvious

Also, are you sure you're not a 13 year old with who's never heard of a t-test in their life?
Yes. Are you sure that you weren't high when you said a random test of 1000 people out of 315 million has a 97% chance of producing a completely accurate statistic?

Blackflag: there is no point in debating the blatantly obvious
Says the person who derides me in the other thread for refusing to engage with him. This IS blatantly obvious if you understand some basic facts about random sampling. You should have learnt this stuff in high school. The formula above is easy senior high school level stuff.

Are you sure that you weren't high
I've never been high before in my life.

when you said a random test of 1000 people out of 315 million has a 97% chance of producing a completely accurate statistic
That's not what I said. Again, you have obviously forgotten high school level maths and can't remember what a margin of error is. No wonder you don't understand what Elo scores are. Those are actually complex by comparison.

admin: This IS blatantly obvious if you understand some basic facts about random sampling.
Replace the word facts with lies and I agree with you.

You should have learnt this stuff in high school. The formula above is easy senior high school level stuff.

Maybe not so easy since your clearly do not understand it yourself. You might no the equations, but you have no idea about the proper application. If for some reason that was the proper application, then I would also make the claim that you got the equation wrong.

I've never been high before in my life
It is possible to get high on idealistic stupidity

Admin, get your head checked out. You have been shouting some wacky stuff lately. This is nowhere close to taking the prize.

Blackflag: Please, enlighten us as to how to properly calculate a margin of error off a univariate statistic in a random sample. I got an A in my university statistics paper - I don't think they usually give those to people who can't do something as simple as a margin of error calculation. But if you think I'm wrong, feel free to humor us.

Actually, don't. Kids, if you're doing your stats homework and want to know how to calculate this, just follow my example above. Assuming it's a normal kind of survey you can't go wrong.

Blackflag: I haven't been shouting. You've been begging me to respond to your various inane ramblings about absolute freedom, and I have. In this case you're attacking one of my stats because you couldn't back up one of yours, with an attack that is mathematically falsifiable.

Please, enlighten us as to how to properly calculate a margin of error off a univariate statistic in a random sample.
I have already taken the time to enlighten you about your error in thinking. Please follow step two in the Stag manual of enlightenment, and look up the truth using google.

I don't think they usually give those to people who can't do something as simple as a margin of error calculation.
If you had to right a paper on proper application, you would of probably got an F.

Admin, if you are never high, perhaps you should consider it.

That could help...

Blackflag: look up the truth using google.
First result:
http://stattrek.com/estimation/margin-of-error.aspx?Tutorial=AP
Agrees with my formula (uses a different arrangement though, my version is easier).

Second result:
http://www.dummies.com/how-to/content/how-to-calculate-the-margin-of-error-for-a-sample0.html
Exact same as my formula

Third result:
http://statistics.about.com/od/Applications/a/Statistics-And-Political-Polls.htm
Exact same as my formula

Fourth result:
https://www.surveymonkey.com/mp/margin-of-error-calculator/
Slightly different, faster & less accurate formula, using this gives a rounded value of 4% instead of 3%

Stag, there isn't a statistician in the world that would tell you that with a .95 CI you'd get a margin of error anywhere near 100% if your sample size is something like 1000. Never anywhere close.

admin: In addition, Google isn't always right. I know this from books, scholarly articles, and teaching from top statistics researchers/professors. The margin of error is approximately 3%, there's no way around it. You are literally the only person in the world who thinks it is close to 100% in this case.

My guess is that you either fundamentally misunderstand the definitions here, or you're trolling.

admin: Did google tell you that a survey of 1000 people out of 315 million is accurate?

Seriously mate, slowly come back to planet earth.