a randy Oliver snip followed by > my comments.. 
Data is often messy--stats help us to make
objective sense of it.  But stats can also be used to reflect one's biases
(which I often see in published papers).  I take all stats with a grain of
salt--I put more trust in my eye, and in comparing replications for
consistency.

>Well sometimes data is messy and sometimes not so much. Largely this has to do with the number of variable you are considering.  In simplistic terms of two variable and two axis then yes trend lines are easy to visualize. With 3 variable any trend is more difficult and with 4 even those with lots of higher math in their background will have problems visualizing the relationship.  Having done more than my fair share of designing structures and building those I will trust the math and the tools to my own eye < fire anyone who trust their eye to run a drainage line.

>Statistics is a very wide field of study and statistical analysis is quite robust and any given stats is designed to handle certain fields or problems (biology, economics, physics, quality control at the end of a production line).  No professional academic would consider using the improper stat since this would decrease any potential of publication to zero at the very first review.  Lots of folks (with very limited understanding of science, science methodology and stats) often demean and devalue stats since they have never personally witness the power in those numbers. Stats becomes the modern day whipping boy of those with limited expertise or experience.

>again it is interesting to see how people who claim to value science display their own bias to demean and devalue the work of other scientist... which just between me and you'ins may have significantly more and deeper understanding of these issues than the person hurling the charge.

Gene in central Texas       

             ***********************************************
The BEE-L mailing list is powered by L-Soft's renowned
LISTSERV(R) list management software.  For more information, go to:
http://www.lsoft.com/LISTSERV-powered.html