How to Fool Yourself—And Others—With Statistics | William M. Briggs
http://wmbriggs.com/blog/?p=2179
---
The use of statistical analysis to draw conclusions from seemingly 
confused data is one of my biggest concerns about the studies we read 
and (try to) believe.  I really wonder how valid much of the statistical 
analysis used to arrive at conclusions we use for bee management 
actually was.  I seldom see the details and often the statistical work 
is mentioned with a quick phrase that sounds authoritative, but does not 
tell us the details.

We often base economically important decisions on the results of 
studies, (although actual commercial beekeepers who live or go broke by 
the results of their beekeeping tend to be highly skeptical about what 
they read, and often, it seems, consider beekeeper coffee clatche 
scuttlebutt more trustworthy).  Personally, I have had a problem 
reconciling my real world experience with what I would have expected 
from some studies I have read.

Often we don't see the data or how the data was selected, then filtered, 
or the distributions and assumptions used.  Often we don't know whether 
the researchers actually understood the requirements for proper use of 
the tools and did the job themselves, or just farmed out the job to an 
expert who did not actually understand the intricacies of bee biology 
and behaviour or how the data was collected and possibly previously 
filtered and/or selected.

I also wonder how qualified reviewers are to critique the statistical 
aspect of studies they read -- or if they are given sufficient data to 
do so.

             ***********************************************
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

Guidelines for posting to BEE-L can be found at:
http://honeybeeworld.com/bee-l/guidelines.htm