Re: RC: Reply from an educated derelict

[Tivers@aol.com]

In a message dated 12/27/99 11:43:55 AM Pacific Standard Time, 
dleblanc@mindspring.com writes:

<< You don't need to prove a negative.  Also, the "games of academics" have
 substantial value in the real world.  I've operated in very different areas
 of research, but it has taught me a little about experimental design.  The
 way you do this is:
 
 1) Define ways to measure benefit
 2) Define ways to measure risks (e.g., tying up, reduction of some defined
 benefit, etc.)
 3) Run an experiment with a test group and a control population.  Use as
 many subjects as you can get your hands on.  Try and eliminate as many
 other factors as possible - train them all the same way, etc.
 
 Now look at the results.  If there were no incidence of negative behaviors,
 then report that.  Note that the smaller the sample size, the greater the
 risk that the population that exhibits a negative behavior was excluded
 from the study by accident. >>


And your estimate of the date that Sarah will get around to this is...

ti


=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
Ridecamp is a service of Endurance Net, http://www.endurance.net.    
Information, Policy, Disclaimer: http://www.endurance.net/RideCamp   
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=