🚀 go-pugleaf

In article <5ea9a1$52q@oolong.memphis.edu>,
Dr. Gordon Kenney <gkenney@memphis.edu> wrote:

>You obviously do not believe in polling or statistical sampling, Steve.  

Of course I do!  When it's done right.  Tell me, Dr., what good is
a poll or survey if it is not representative of the group you are
trying to measure?

>If the poll randomly selected respondents, then one would expect that a 
>percentage relatively equivalent to the percentage obtained from the 
>sample of 1,000 to be in favor of medicinal cannabis use.  1,000 is a 

If the poll randomly selected respondents.  I'm claiming we have no
proof, and good reason to doubt they were truely randomly selected.
If you're trying to get a voice of all America, it better represent
all of America.  Have you ever traveled to different parts of
the country Dr.?   If so, you'd realize that certain areas contain
vastly different opinions.   All you had to do was watch the Presidential
election on TV to realize this.

>pretty good social science sample, if you ask me.  I don't know what you 
>study at OSU (and I certainly hope you do not represent the FACULTY 
>there -- I would expect that they would have more statistical 

You're jumping the gun a little soon aren't you Dr.?   One would think
that with all of your acquired knowledge and degrees that you'd need
a little more proof before you put your stamp of approval on something.
Having said that, I hope you don't represent the FACULTY where you're
at either.

>knowledge), but when pharmaceutical companies "test" prescription drugs, 
>they do not select a sample of 260 million (not 250 million -- the USA 

I honestly don't know.  Is Lake... Research a pharmaceutical company?
How long have they been in business?  What is their track record?
How can we trust them?  Has their study been verified?   If not,
why should we believe them?   THEY WANT TO PUSH DRUGS.  --> What
do you _think_ the result of their survey will be?

>has grown since 1990) Americans to serve as subjects.  That would be 
>WAAAAYYYY too expensive.  A random, double-blind sample (perhaps of 100 
>people) does the trick nicely -- at a fraction of the cost.  After all, 

"Trick" is the keyword.  "fraction of the cost" is the accomplice.

>we don't elect our politicians by asking EVERY citizen to vote (we'd 
>like that to happen).  We are content with a (nonrandom) sample of about 
>30% of them.  If you haven't registered for Fall 1997 classes yet, might 
>I suggest Statistics 101...
>
>-- Gordon

Oh come on Gordon!   We _don't_ ask every citizen to vote?  Where
have you been pal??   In a cave?   We, and all politicians certainly
do ask every citizen to vote, and we hope that they will.

And our politicians are not elected necessarily by the direction national
popular opinion. -only those who vote.  By comparing this to our
supposed random survey above, you're comparing apples to oranges.

I would strongly suggest Statistics 101 for you, Gordon.  I'm plenty
knowledgable about statistics.  Plus I'm graduating next quarter with
a B.S. in Mechanical Engineering.



Steve