My stats prof. had actually been a cowboy before becoming a mathmetician!
He had some great "life on the range" stories, but I digress...  The neatest
thing he had was a video tape (he had taken years to create--some were even
recorded originally in Beta!) of commercials that had used "statistics" to
support a position.  A few were actually marketing a product, but a goodly
number were political ads or "PSA"s.  The scary thing was that a study at
Johns Hopkins demonstrated that a certain percentage of the population at
large will beleive virtually anything if you:  1) Put it in a 'respectable'
media source (meaning, anything except those media outlets that were OBVIOUS
in their approach, such as "National Enquirer", "Globe", etc.).   2) Repeat
the numbers often (e.g., "statistics" quoted during political commercials
during late October).  What was worse was that the study indicated that even
if something was a blatant lie, and the source was forced to make a
retraction, a certain percentage of viewers/readers would STILL BELIEVE the
original material!

I loved that class!  One of our assignments was to bring in a print ad every
class.   The ad needed to use stats in "grey" fashion.  After the first
night of class I thought I was going to be spending all my free time
scouring magazines looking for these ads.  Heck, I think I found several in
the first magazine I picked up.

One classic was the toothpaste ad that said, "four out of five dentists
recommend brand xxxx."  If I remember correctly, a survey had been sent to a
small number of dentists with a question something along the lines of "Would
you recommend your patients use brand xxxx or plain water when brushing."
That's not quite right, but the gist of the question was that it was almost
impossible for the dentist to recommend anything but brand xxxx.

-Kyle

-----Original Message-----
From: Kathy Salkin [mailto:[log in to unmask]]
Sent: Tuesday, October 29, 2002 11:31 AM
To: [log in to unmask]
Subject: Re: another shooting


LOL I'm falling out of my chair, laughing... Yes, my grad school stats prof
used to do the same thing to us, pointing out inane cause and effects
supposedly bolstered by high correlation coefficients.  You're right of
course, numbers can be manipulated to prove any point you wish.

Kat