> reverse correlation between honey yield and pollination fees

This is an example of what seems "data dredging" - past data is analyzed, and a correlation is found, and assumed to be valid.
Very very few such correlations stand up to the test of seeing if they have predictive power and are valid correlations.
Most correlations turn out to be pure luck/chance, as many who have played the stock market have learned the hard way.

But even if the correlation holds up, we have the issue of causality - any beekeeper could observe that both trends might well have been actually "caused" by a 3rd factor, varroa, which both hurt honey crops, and made it harder to provide strong colonies for (almond) pollination, hence driving up the price.
But there are also other well-known factors that clearly contributed here, such as the steady increase in producing almond acres (433,000 in 1995 up to 615,000 in 2007 when I stopped tracking it).

I have attached another negative correlation based upon perfectly accurate past data - as the number of pirates has decreased, global average temperatures have increased.
This illustrates how dangerous a large dataset and a statistics package can be in the hands of the easily amused.



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