A LOT of time was (and is) spend on question design of the COLOSS questionnaires and a lot of things can be said about variable biologists making up variable questions on animals that are basically livestock and so can only be targeted through their variable human keepers. A good survey design course at the social sciences would really benefit things.

In the national survey we find a couple of worthwile questions to ask. What is your location (on postal code scale), do you know if your bees foraged on Tilia, Salix etc., a matrix in which a beek can tick of the months and substance for varroa treatment, and questions on things like comb replacement (how large % replaced yearly).

The statistics to analyse the data are pretty well developped, took me about 3 years, but my (ex)boss is slow with publication for some reasons. I can't say anything about it really as we want to have the methodology scoop.

Currently I am re-analyzing and expanding the PLOS imidacloprid~fauna abundance piece. Animals that are not associated with humans have a bigger problem here. And, for a change, the data on this are much much better. The human variability is out, beek and their associated researchers are just so variable. For example, with the bee data overdispersion (variance~mean relationship not standard) is caused by comparing small and big beek in small data. The abundance data are also overdispersed but as the data are much larger this is way less of a problem, things are less rare and dispersion tends to be related to actual field phenomena. Not that beekeepers are not a field phenomenon but there is not much to measure about them.

If I had to bet on good additional questions for bee survey I would bet on things like age/income of beek, religion/education of beek. When playing with this for fun we find gender effects, women beek have lower loss, but they are rare and prolly sex of beek is associated with habitat. More female beek in urban areas with more flowers.

Lennard

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