Accuracy and understanding Statistics.

[~candice~]

That "50% of Christians divorce" stat has always been skewed.

First off, lots of people claim to be Christian when they actually are not. Lots of folks believe that just because they were baptised as a baby, or they go to Church, or the come from a Christian family, or they simply believe in God, it makes them a Christian.

Also, it doesn't take into account when a Christian is divorced by a non-Christian.

Or if they got divorced before they were a Christian!

This is easily determined; just check the kurtosis.

Fond of the fourth moment?

Not especially, but it does help to know when the data is wonky.

You just made my day

[~candice~]

In college I took a Statistics course. The only thing I learned that I remember to this day is that statistics can be skewed to back up any point of view you want. That was the point when I stopped paying attention to statistics.

That isn't true at all

If you did alittle research you would see that what she said is true.

I have a phd in statistics... just exactly what research are you talking about?

A phd......wow. I don't have one and it took me all of 2 minutes to find this and this.

LOL.. the first link came up with an answer in the negative, the second said exactly what I said. A quote by Mark Twain doesn't a truth make.

A workman never blames his tools . It's just a tool, use or misuse at your own risk.

[Guest man]

LOL.. the first link came up with an answer in the negative, the second said exactly what I said. A quote by Mark Twain doesn't a truth make.

A workman never blames his tools . It's just a tool, use or misuse at your own risk.

You sound confused. You first say that statistics can't be skewed and then you say they can.

In college I took a Statistics course. The only thing I learned that I remember to this day is that statistics can be skewed to back up any point of view you want. That was the point when I stopped paying attention to statistics.

That isn't true at all

Sounds as if people misunderstand... statistics is just a tool. Actually it is just a summary of sample data, presumed to resemble the true value of a parameter from the population of interest. It's a bit like a type writer... the statistic is only a tool, not the author. It depends on who is pushing the keys and the error is in the interpretation, not the statistics itself.

The potato example, is a question of (internal) validity. The statistic itself is fine, but it doesn't measure what was claimed to be measuring. The most common way to lie is to take a sample from population A and try and apply it to population B.

You can't force data to lie. All you can do is misrepresent the extent to which your conclusions from that data, apply to other populations. That's not a problem with the tool statistics) but with the author / story teller.

Turns out there are two kinds of people in the world ... liars, bad liars, and people who misuse statistics .

[~candice~]

LOL.. the first link came up with an answer in the negative, the second said exactly what I said. A quote by Mark Twain doesn't a truth make.

A workman never blames his tools . It's just a tool, use or misuse at your own risk.

You sound confused. You first say that statistics can't be skewed and then you say they can.

In college I took a Statistics course. The only thing I learned that I remember to this day is that statistics can be skewed to back up any point of view you want. That was the point when I stopped paying attention to statistics.

That isn't true at all

Sounds as if people misunderstand... statistics is just a tool. Actually it is just a summary of sample data, presumed to resemble the true value of a parameter from the population of interest. It's a bit like a type writer... the statistic is only a tool, not the author. It depends on who is pushing the keys and the error is in the interpretation, not the statistics itself.

The potato example, is a question of (internal) validity. The statistic itself is fine, but it doesn't measure what was claimed to be measuring. The most common way to lie is to take a sample from population A and try and apply it to population B.

You can't force data to lie. All you can do is misrepresent the extent to which your conclusions from that data, apply to other populations. That's not a problem with the tool statistics) but with the author / story teller.

Turns out there are two kinds of people in the world ... liars, bad liars, and people who misuse statistics .

Sounds like you are trying to generate an argument when none is there...

Statistics are just numbers and can't lie, the lie is in how they are presented to support something they have no business supporting. But by all means, think what you want, you are the expert after all...

[Guest man]

Sounds like you are trying to generate an argument when none is there...

Not really. It was just an observation.

Statistics are just numbers and can't lie, the lie is in how they are presented to support something they have no business supporting.

That's what everyone's been saying........................well almost everyone.

But by all means, think what you want, you are the expert after all...

Actually, you're the one with the phd.

[Guest]

.... trying to generate an argument.....

Statistics are just numbers and can't lie, the lie is in how they are presented to support something they have no business supporting. But by all means, think what you want, you are the expert after all.....

OK!

Generating.....

99.999% Of Apples Found On Horse Trails Smell (Good Math - One Population)

Therefore 99.999% Of Apples Are Smelly (Same Math - Another Population)

So Do I Get That Political Advertising Writing Job Please Please Huh Huh

[~candice~]

Well done Joe now you get to be a politician

[Guest]

Well done Joe now you get to be a politician

Thank You!

I Understand I'm On The Ticket For President/Vice President

Just Need To Toss A Few Road Apples Into The Mix And I'm Sure To "Slide" Into Victory! :24:

[~candice~]

[Believer1997]

As a statistician, based upon the particular study, random sampling, linear statistics and curve for which you would discard and/or include some data... and presentation... statistics can be distorted as much as 50% of the time. Sound high? Well... it is. Thus, you have half a chance of finding out if something is accurate or not, which would behoove one to peruse a number of other similar studies to determine if you can get within a percentile that may give you a loose interpretation of what is being studied and how accurate the stat is.