I do not understand the regular bashing that the BIP gets for simply being a
survey and plainly labeled as such.
Let me just list off the top of my head some terms that I hear the
statistics folks toss around when issues of "selection bias" are raised:
"Heckman Correction", "Chamberlain's conditional fixed effects approach",
"Berkson's fallacy", the list goes on and on. But the easier answer is to
participate, and encourage others to participate, so as to get a broader
data set at the start.
But the more basic problem is that an accurate assessment of "cause of
death" for a hive is a subject fraught with rationalization, excuses, and
self-justification, even when clear evidence makes a rare appearance, and
the beekeeper is experienced enough to notice the evidence.
What is needed is a metric that is independent of the beekeeper's need to
reassure himself/herself that the cause of death was not neglect. I sell a
relatively inexpensive tool that can be used to track that metric. If one
merely weighs one's hives, and subtracts and adds the weight of what one has
added or removed, one ends up with the weight of the colony stores over the
season, as bees and brood weigh essentially nothing in the grand scheme of
things.
The colony weight alone can reveal the significant events that result in
failure to survive, and hence, be a better basis for evaluating the "cause
of death". Will a 10,000-colony operation weigh each and every hive every
week? Nope, for that scale of operation, the best you could do would be to
build scales into their forklifts, and weigh "by the pallet". You'd also
need metadata on what the colonies were doing each week.
But eliminating "opinion" from the survey would do wonders, as colonies tend
to dwindle and lose weight at different times of the year as a result of
different problems.
***********************************************
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
|
