[Prev][Next][Index][Thread]

Re: Kenny fails YOUR stats!



In article <40dcqa$ss7@news-e1a.megaweb.com>,
KENNY MORSE <kennystar@megaweb.com> wrote:
>>> who absolutey reflect a cross section of California drivers who were
>>      ^^^^^^^^^
>>
>>This is called "proof by vigorous assertion".  You'll have to do 
>>better than that.  
>Nope...I don't have to. In YOUR business....when people
>try to tell you YOUR area of expertise, I bet YOU are vigourous
>too when you disagree.  I will always assert that which I know to
>be true.  So will you.  We can disagree.. That is what
>places like this are for.
>
>Sorry you missed it. I answered each class.
>>Second, the ones who are "caught" more often (and hence are more
>>likely to choose to go to traffic school) tend to be the more
>>dangerous drivers
>
>Your premise is faulty...."caught more often?"
>Who says they get caught more often just because they are in
>traffic school???
>

This is one of your points you keep making, Kenny, that is not a matter
of contention because we disagree about why people speed, or what the
law should be, what people's behavior is under different road conditions,
or the behavior of cops. This particular point is under contention
because you do not seem to understand something very essential in
polling and sampling.

Here is the fact: the people you see in your traffic classes are *not*
a cross-section of all California drivers.

Here is why: people who get more tickets will tend to show up in your
classes more. The average person in your classes will have more 
tickets than your average California driver.

Let me illustrate. I don't know what the rate at which people go
to traffic school as opposed to just paying their tickets, but
for the sake of argument, let's say that 25% of all tickets get
their recipient to go to traffic school rather than pay their 
tickets. Let us also say that an equal number of drivers in a given
period of time receive 0,1,2,3, and 4 tickets. Let's say that there
are ten million drivers in the state.

Note that the actual numbers and distributions are unimportant, but
they are just being used to illustrate a point.

OK, back to the illustration. Given these figures, we can calculate
some things about the drivers. What is the average number of tickets
a driver receives in this time period (we'll call it a year)? Two.
There are twenty million total tickets issued. Two million people
each receive 0,1,2,3, and 4 tickets. There are 0 trips to traffic
school from people with 0 tickets. There are 500,000 trips to traffic
school from people with 1 ticket. There are 1,000,000 trips to
traffic school from people with 2 tickets. There are 1,500,000
trips to traffic school from people with 3 tickets. There are
2,000,000 trips to traffic school from people with 4 tickets.
This makes a total of 5,000,000 trips to traffic school. However,
40% of those trips were from people with 4 tickets, 30% of those
trips were from people with 3 tickets, 20% of these trips were from
people with 2 tickets, 10% of these trips were from people with
1 ticket, and 0% of the trips are from people with 0 tickets.
This is *not* a representative cross-section of California drivers!
A representative profile of drivers in this theoretical state would
consist of equal 20% parts people that received 0,1,2,3,and 4
tickets. The average number of tickets received by a person in
a traffic school class is ((.40*4) + (.30*3) + (.20*2) + (.10*1) +
(0*0)) = 3, which is 50% over this theoretical state's average.

The numbers are unimportant, but I hope you see the point I'm
illustrating. People who get caught more are more likely to
show up in your class. Maybe they're not more likely to show up
*for a particular ticket*, but because they receive more tickets,
they are more likely to go to traffic school. Taking extreme
cases, someone who never got tickets would never go to your class,
and some schmuck who got a ticket a week would be pretty likely 
to show up in traffic school.

And I hope we can agree that someone who gets 4 tickets over
a period of time is most likely a very different driver than
one who gets no tickets during that same period, unless you
think tickets are handed out in a purely random fashion. Surely
chance does play a (perhaps major) factor in whether someone
receives a ticket, but averaged out over a large population,
those who receive more tickets will be a different group of
people than those who do not receive a large amount of tickets.



-- 
Adam Villani
addam@cco.caltech.edu
http://www.ugcs.caltech.edu/~addam
"I was in this prematurely air-conditioned supermarket..."


References: