JFern's "Statistics"

[○∙◄☻¥tπ[╪AV┼cVê└]

You're asking for a weighted random sample. That would be less random then a random sample. If you knew you had the true weight (let's say 35% Dem, 35% GOP, 30% Indepdendent, or something, I made those numbers up), then you can calculate the MOE on each part equally, and find the combined MOE weigted MOE, by finding the MOE of each subsample, and adding them using the Pythagorean formula for adding standard deviations.  Then, that would give you a smaller MOE, but it's not what I was talking about because didn't want to get into complicated matters like weighted non-random polls.

No, weighting is not the same thing as I am talking about.  I am talking about validity checks.

Simple question:  Is a poll of 1000 registered Democrats livining in Boston representative of the general electorate?

Random samples can and  blow up.  They can pull in far too many members of any group.  We know from the census that the US population has 12.9% African Americans.  Let's say that If a poll gets 20% African American respondednts that is too much for the nation.  We can either weight it down to a more reasonable amount, increase our sample size by interviewing people of other ethnic groups, toss the poll out and start over or, if we are irresponsible, release it and pretend it is a propre representative sample.

If we get a poll that is 14% African American we can weight it down, or leave it be since it is close enough.

A poll can be picture perfect, spot on for every measurable demographic and perfect in every technical aspect with questions that are perfectly neutral and readers who show no prejudice one way or another and still be completely wrong.

There's nothing wrong with a completely random sample, you get the MOE we were talking about.

Now, validity checks are good if you know the breakdown of the population very accurately, and
1. You think your sample has some systematic bias you'd like to correct
2. You'd like somewhat smaller MOE. For example if I have a weighted poll, with the correct weight, and 90% of Dems choose Kerry, and 90% of Republicans choose Bush, this makes for a smaller MOE, since MOE is largest when we're near a tie.


Anyways, this wasn't relevant to my argument with J.J.

[J. J.]

A poll can be picture perfect, spot on for every measurable demographic and perfect in every technical aspect with questions that are perfectly neutral and readers who show no prejudice one way or another and still be completely wrong.

There's nothing wrong with a completely random sample, you get the MOE we were talking about.

Now, validity checks are good if you know the breakdown of the population very accurately, and
1. You think your sample has some systematic bias you'd like to correct
2. You'd like somewhat smaller MOE. For example if I have a weighted poll, with the correct weight, and 90% of Dems choose Kerry, and 90% of Republicans choose Bush, this makes for a smaller MOE, since MOE is largest when we're near a tie.


Anyways, this wasn't relevant to my argument with J.J.

Well Tedrick understands it; this is what we are arguing about.  Tedrick has it completely right in that statememnet.

Let's try it again:

1.  The score, or result that a candidate gets, is within a MOE.  That means if the poll is accurate the candidate "real" score is within the MOE.

2.  The same poll, will not be accurate 5% of the time.  The MOE does not effect this at all.  This is what I have been referring to since the start of the thread.

What part of these two statements do you have a conceptual problem with? 

[○∙◄☻¥tπ[╪AV┼cVê└]

You didn't respond to my points. You lose.

[J. J.]

You didn't respond to my points. You lose.

JFRAUD all of of your questions were answered initially.  You just were not bright enough to understand the answers.  You did not ask if this was within the MOE or within the confidence interval (which are one and the same).  You post drivel, like that MOE = 1.96 Standard Deviation, and then offer no support.  Where is it?

What part don't you understand about these to statements:

1.  The score, or result that a candidate gets, is within a MOE.  That means if the poll is accurate the candidate "real" score is within the MOE.

2.  The same poll, will not be accurate 5% of the time.  The MOE does not effect this at all.  This is what I have been referring to since the start of the thread.

Tedrick understands, why can you?

[○∙◄☻¥tπ[╪AV┼cVê└]

Let me post this again, since you were too stupid to respond to it the first time:

http://americanresearchgroup.com/moe2.shtml

That website uses the term statistically significant lead that you argue against. In addition, if you view the source of the webpage, you'll see that it uses 1.96.

The calculator on that webpage uses 1.96 standard deviations to find what is a statistically significant lead, and with a sample of 1000, even 54%-46% is a statistically significant lead.

I've mentioned this repeatedly, and it's clear that you're porposely ignoring this because it proves that you're an intellectually dishonest lying hypocrite.

[○∙◄☻¥tπ[╪AV┼cVê└]

Note to lying sack of sh**t J. Idiot, I'm not responding again until you address those points.

[J. J.]

Note to lying sack of sh**t J. Idiot, I'm not responding again until you address those points.

Here is the response:

1.  The score, or result that a candidate gets, is within a MOE.  That means if the poll is accurate the candidate "real" score is within the MOE.

2.  The same poll, will not be accurate 5% of the time.  The MOE does not effect this at all.  This is what I have been referring to since the start of the thread.


Do you follow this so far?

Now, you never asked me if the result was outside the MOE.  My answer would have been probably yes.  That however does mean that the actual score that the candidate has is within the MOE.

You asked about the significance of the poll.  As Tedrick pointed out, you can still get a bad, or unrepresentive sample.  At a 95% confidence level, this poll will be correct 19 out of 20 times, but there is no way to know if this poll is one of the 19 from the numbers within the poll.  Even a well constructed and well executed poll will be wrong 1 in 20 times at the 95% confidence level.  That what I said on page one:

The thing is, even if the pollster does everything right, he's still going to get one absolutely wrong result out of the margin of error in X amount of polls.  Now, with a sample size of 1000, X will be lower than if the sample size is 300, but it's still going to happen.

Actually, sample size will not change the Confidence Level (CL), but it will affect the MOE.  Whatever the sample size and whatever the MOE, at the 95% CL, one in twenty polls will not show an accurate result even within the MOE. 

If the total population is 120,000,000 voters, 1000 voters are polled, and Kerry has 70%, the MOE is +/- 2.84.  So according to the poll, Kerry has something between 67.16% and 72.84%, according to the poll, at a 95% confidence level. 

Do you follow that so far?

Now, what is the chance that the poll, like most of your posts, is total garbage?  5%  Why, because poll itself, no matter what score is, is still going be wrong 1 in 20 times.  There is no way to determine if this poll is garbage based on the interntal numbers.  While the MOE would change, the poll is always going to be wrong 1 in 20 times; the real result will NOT be within the MOE.

You posted this gobbledegook:

I pointed out you were wrong about saying I was wrong about MOE=1.96 standard deviations - no reply
I asked you why they say that a poll of 1000 says

MOE does not equal 1.96 Standard Deviation (SD).  MOE does not equal any Standard Deviations (SD).  SD basically is refered to as the Confidence Level; a 95% confidence level is two SD from the median.

MOE is determined by several things, what confidence level you are looking at, what the sample size is and what the score is.

Do you follow this so far?

Now, you (and anyone else interested) can test this online, from a site you posted:

http://www.surveysystem.com/sscalc.htm

Do you follow that?  It's possible too determine through a test if I'm correct.

In the above example, with a population of 120,000,000, and a sample size of 1000, and a 95% confidence interval, Kerry had 70%.  If you plug those numbers into the Find Confidence Interval calculator (the lower one), it will yield a MOE or "Confidence Interval" of 2.84.

Now, change the Kerry result and calculate it again.  You will get an MOE of 3.1.  The SD or Confidence Level, has not changed.  Even with this change, there is still a 5% chance the poll is wrong.

It's a simple test, so even JFRAUD the Dense should be to follow that.  He really should have tried it prior to posting.

[○∙◄☻¥tπ[╪AV┼cVê└]

You didn't address this. You lose. Game g over.
Since you are incapable of responding to this post that I've posted several times, I shall despise you as the biggest liar on this entire forum. Even nomorelies is more honest than you.

I've dumbed down my argument so that even a moron like you should be able to realize you're wrong, and you don't respond. I didn't ask you to repeat what you've already said in other posts that doesn't address this, I asked you to respond to the very specific stuff mentioned here.

http://americanresearchgroup.com/moe2.shtml

That website uses the term statistically significant lead that you argue against. In addition, if you view the source of the webpage, you'll see that it uses 1.96.

[J. J.]

You didn't address this. You lose. Game g over.
Since you are incapable of responding to this post that I've posted several times, I shall despise you as the biggest liar on this entire forum. Even nomorelies is more honest than you.

http://americanresearchgroup.com/moe2.shtml

That website uses the term statistically significant lead that you argue against. In addition, if you view the source of the webpage, you'll see that it uses 1.96.

JFRAUD the Dense they are referring to the MOE.  Had you asked me if  this was outside of the MOE, I would have said yes.  You were too stupid to ask.

But even being outside of the MOE, the poll still has the chance of being garbage.

[○∙◄☻¥tπ[╪AV┼cVê└]

You didn't address this. You lose. Game g over.
Since you are incapable of responding to this post that I've posted several times, I shall despise you as the biggest liar on this entire forum. Even nomorelies is more honest than you.

http://americanresearchgroup.com/moe2.shtml

That website uses the term statistically significant lead that you argue against. In addition, if you view the source of the webpage, you'll see that it uses 1.96.

JFRAUD the Dense they are referring to the MOE.  Had you asked me if  this was outside of the MOE, I would have said yes.  You were too stupid to ask.

But even being outside of the MOE, the poll still has the chance of being garbage.

I made two points

1. They use 1.96 standard deviations to calculate MOE (they assume large samples)
2. If you're outside the MOE, they say you have a statistically significant lead

I haven't seen you respond to either of those points.

[J. J.]

I made two points

1. They use 1.96 standard deviations to calculate MOE (they assume large samples)

JFRAUD here is exactly what you said:

I pointed out you were wrong about saying I was wrong about MOE=1.96 standard deviations - no reply
I asked you why they say that a poll of 1000 says

Now, perhaps you made a mistake in posting it, like so many other things, but I have responded, accurately.  Do you still think that "MOE=1.96 standard deviations?" 

I unfortunately cannot respond things you didn't say.

You must be logged in to read this quote.

Again, I have answered.  It is outside of the MOE, but we really cannot determine it the poll is accurate, based on the results of the poll.

I actually answered both of these on the first page:

The lead itself would not be statically valid, as you would have to look at the poll itself.  You are equating the result with the validity of the poll. 

[○∙◄☻¥tπ[╪AV┼cVê└]

I made two points

1. They use 1.96 standard deviations to calculate MOE (they assume large samples)

JFRAUD here is exactly what you said:

I pointed out you were wrong about saying I was wrong about MOE=1.96 standard deviations - no reply
I asked you why they say that a poll of 1000 says

Now, perhaps you made a mistake in posting it, like so many other things, but I have responded, accurately.  Do you still think that "MOE=1.96 standard deviations?" 

I unfortunately cannot respond things you didn't say.

You must be logged in to read this quote.

Again, I have answered.  It is outside of the MOE, but we really cannot determine it the poll is accurate.

I actually answered both of these on the first page:

The lead itself would not be statically valid, as you would have to look at the poll itself.  You are equating the result with the validity of the poll. 

1. You say that I'm wrong for saying that for large samples, and 95% confidence intervals, MOE=1.96 standard deviations. You said the correct value is 2. This is another example of you being wrong.

2. We're assuming a perfectly random sample of 1000 people. It's automatically statistically valid. We merely choose the confidence level for which we consider it to be statistically significant. I see you don't like the term statistically significant lead, but it's used in the page I linked to. That page assumes nothing about the poll, and tells you all you need to know about whether there's a statistically significant lead.

You still haven't addressed these you pathetic fraud. Even when it's obvious that you're completely ing wrong, you try to claim I'm wrong. You are the most pathetic of the pathetic.

[○∙◄☻¥tπ[╪AV┼cVê└]

I think if you have any intellectual reasoning, you should be able able to determine that you're wrong, now that I dumbed my argument down enough for you. With the clear evidence that you're wrong and I'm right, your continued insistance that you're right and I'm wrong shows that you are not capable of the most basic types of intellectual arguments. I have admitted when I've been wrong on this forum. You have been wrong far more often, and have not admitted when you're wrong. I will not waste my time arguing with a complete fraud such as yourself.

[○∙◄☻¥tπ[╪AV┼cVê└]

I conclude that in J.J world the following are true (1 is more important).

1. If a coin is tossed 1000 times, and 940 heads and 60 tails come up, that there isn't statistically significantly more heads than tails, and you can not conclude that the coin is statistically significantly different from a fair coin.

BTW for a false positive probability p, you can choose 5%, 0.1%, or 10^-200.

2. For a large enough statistical sample (which would follow the normal distribution), you think that the width of the 95% confidence interval is exactly 2 standard deviations (because you heard of some 68-95-99.7 rule), and not approximately 1.96 standard deviations.

Anyone knowledgeable about statistics should be able to conclude that you're a complete fraud for defending those positions in face of overwhemling evidence.

[J. J.]

1. You say that I'm wrong for saying that for large samples, and 95% confidence intervals, MOE=1.96 standard deviations. You said the correct value is 2. This is another example of you being wrong.

JFRAUD the Dense has just revealed is total ignorance of statistics again.

Yes, you moron, MOE does not equal any standard deviation; I have no idea why you continue to post it.  You can try it experimentally on the calculator.  Keep the 95% confidence level and change the result from 70% to 50%; the MOE changes.  Even you should be able to do it!

BTW:  Here is the link:

http://www.surveysystem.com/sscalc.htm

You, or anyone else can demonstrate the effect.

You must be logged in to read this quote.

You do not get an automatically statistically valid sample size, unless you poll the entire population (and yes, in the real world there could still be problems).  The page says at the 95% confidence level, and that has been posted on this thread.

Here is what the site say:

You must be logged in to read this quote.

 Emphasis added.

JFRAUD the Dense, they are referring to MOE.  If you would have asked if this was outside of the MOE, I would have answered yes.  That has nothing to do with the poll being accurate.  That is what I have been referring to.

Interestingly, in your first question you said:

This is relevant, because if you answer yes, you are contradicting what you argued for months, not admitting defeat, so to be consistant, you have to answer no, a lead of 94% to 6% in a standard opinion poll is not a statistically significant lead.

I'm assuming that this refers to the Global warming thread.  

In that one, MOE would not come into play at all.  The sample size was 100%.  The MOE or confidence interval would be ZERO.  If you polled 100% of the voters, the MOE would be zero as well.

I'm so happy that you linked to this thread in you sig.  It so clearly illustrates your lack of understanding, especially of statistics.

[J. J.]

JFRAUD, simply run the experiment I suggest and report the results.

[○∙◄☻¥tπ[╪AV┼cVê└]

You ing idiot.

Same two points again that you haven't addressed properly.

1. Quoting other random stuff isn't changing the basic problem. Let it be a random sample of 1000 voters with Kerry at 94% and Bush at 6%, or a coin flipped 1000 times with 940 heads and 60 tails. It's the same problem. Now, you take this poll, and go to the ARG page that you like to ignore. That page tells you that even  a 54%-46% lead is a statistically significant leadwith a sample of 1000, so of course a 94%-6% lead is.

2. For large samples, and even larger populations, the normal distribution gives a very good approximation, and for the normal distribution the 95% confidence interval is 1.96 standard deviations in either direction. Of course the MOE varies as the percentage, because if you knew basic statistics, you'd know that the standard deviation is sqrt(p*(1-p)/n), which obviously varies with the percentage.

Why am I bothering? It's obvious at this point that you'd argue 2+2=5 just to avoid admitting you were ever wrong.

[○∙◄☻¥tπ[╪AV┼cVê└]

JFRAUD, simply run the experiment I suggest and report the results.

Yes, we've determined that MOE varies with the percentage in the poll as everyone who remembers basic freshman statistics knows. Do you have a ing point, you ing moron?

[○∙◄☻¥tπ[╪AV┼cVê└]

Do you understand anything about statistics?


Method 1:


1. Set up a null hypothesis (in this case that Kerry and Bush are tied at 50%, or that the coin is fair)

2. Choose a small enough false reject probability p

3. Collect a sample of data

4. See if the sample you collected is statistically different from the null hypothesi (in this case that Kerry or heads has a statistically significant lead)

Method 2:
Similiar, do step 1, and then step 3, and in step 4 calculate the false reject probability p that it is statistically significant at. In this case p< 10^-200.



Time to go back to freshman statistics.

[○∙◄☻¥tπ[╪AV┼cVê└]

How can you live with yourself for being such a ing fraud?

[J. J.]

You g idiot.

Same two points again that you haven't addressed properly.

1. Quoting other random stuff isn't changing the basic problem. Let it be a random sample of 1000 voters with Kerry at 94% and Bush at 6%, or a coin flipped 1000 times with 940 heads and 60 tails. It's the same problem. Now, you take this poll, and go to the ARG page that you like to ignore. That page tells you that even  a 54%-46% lead is a statistically significant leadwith a sample of 1000, so of course a 94%-6% lead is.

Wrong again, JFRAUD.  The 1000 is a sample, in the polling case.  And at the confidence level, it will always be wrong a percentage of the time; that is where the "confidence level" comes in.  As has pointed out the page refers to a Confidence Level of 95%.

2. For large samples, and even larger populations, the normal distribution gives a very good approximation, and for the normal distribution the 95% confidence interval is 1.96 standard deviations in either direction. Of course the MOE varies as the percentage, because if you knew basic statistics, you'd know that the standard deviation is sqrt(p*(1-p)/n), which obviously varies with the percentage.

Gee JFRAUD. I wish I'd said that.  Oh, wait a minute, I did:

MOE does not equal 1.96 Standard Deviation (SD).  MOE does not equal any Standard Deviations (SD).  SD basically is refered to as the Confidence Level; a 95% confidence level is two SD from the median.

MOE is determined by several things, what confidence level you are looking at, what the sample size is and what the score is.

At least you are moving in the right direction, finally.

I suspect that you are "bothering" because you think that this affects your credibility.  Don't worry, this isn't as bad claiming that something that happened in July of 1933 triggered something in March of 1933.  :-)  That was last week.

Of course, you could also treat this as a learning experience, in all seriousness.  I suspect you tried the experiment, and that has caused you modify your stance a bit (okay a lot).  That's good.  Frankly, untill I tried it, I didn't realize how many things affected MOE, and it's clear you didn't either.

Now, if you really tried to read and think before you link, you might get further, or try something really radical, like asking questions.  Or you stay the same old arrogant idiot, make mistakes, and get called on them.  

Personally, I'd rather see you make more accurate posts.

[J. J.]

How can you live with yourself for being such a g fraud?

My apologiez, it was too much to ask for you to attempt to post accurately and cogently.

JFRAUD he remains.

[○∙◄☻¥tπ[╪AV┼cVê└]

I notice you didn't respond to my post on the steps you make to get statistical significance.

The percentage of the time you falsely conclude that there is statisitical significance of something that isn't true is at most p of the time, which is at most 1 in 20 for p=5%. Just because it happens to be the 1 in 20 poll where I falsely conclude statistical significance when it's not true doesn't mean that the poll isn't statisically significant.

About the approximately 1.96 standard deviations for the 95% confidence interval for the normal distribution, if you had  1. looked at the source for that ARG calculator or 2. looked at where I calculated that 2 standard deviations gives 95.44%, using the normal distribution density function of  1/sqrt(2*Pi) * e^(-x^2/2), you'd realize you were wrong. 

I never said that events in July 1933 happened before events in March 1933, you g sh**thead.

Now you still haven't responded to discussion of basic hypothesis testing ffrom freshman statistics. I'd suggest you take that course, since you don't know g sh**t about statistics.

[○∙◄☻¥tπ[╪AV┼cVê└]

This is ing hopeless. What would it take for you to ever admit that you're completely wrong?

[○∙◄☻¥tπ[╪AV┼cVê└]

I notice you complete lying sack of worthless dog sh**t didn't respond to this.

Do you understand anything about statistics?


Method 1:


1. Set up a null hypothesis (in this case that Kerry and Bush are tied at 50%, or that the coin is fair)

2. Choose a small enough false reject probability p

3. Collect a sample of data

4. See if the sample you collected is statistically different from the null hypothesi (in this case that Kerry or heads has a statistically significant lead)

Method 2:
Similiar, do step 1, and then step 3, and in step 4 calculate the false reject probability p that it is statistically significant at. In this case p< 10^-200.



Time to go back to freshman statistics.