> Where is the part where you evaluate the analysis? 

The term "preliminary" is key here.
Without at least a preprint, one cannot attempt to evaluate, as all one has
is an announcement about a paper yet to be published.

>  You started off well and ended irreverently, but somehow the sandwich has
no meat in it.

The question I was addressing was "how [is] our thinking is wrong or
correct?"
The meat is the math. 
With data, the meat is always the math.
I gave several examples of how they could statistically account for
non-responding operations/beekeepers.
An attempt to critique "a survey" in absence of knowledge of the specific
statistical methods used is mostly futile.
So, critique of a survey is inherently a discussion of statistical methods.

If I have any "analysis" at this point, it is this:

In other fields of science, this sort of "pre-announcement" is usually done
via posting a pre-print of the full paper to ArXiv or one of the other
preprint servers, with a note that the paper had been submitted to journal
X, and making clear if the paper is likely be under embargo or not.
Now, if the journal is one with an embargo, it is much easier to comply with
the terms of the embargo if one does not "privately" distribute the paper
under potential embargo to the several thousand people in hopes that, as
Phil Ochs sang " It wouldn't interest anybody - outside of a small circle of
friends".

So, I am left with an ethical dilemma - do I freely discuss the claims
before they are published, and risk hurting the chances of the results being
published in a journal with an embargo policy?
Or do I simply observe that the term "preliminary" is self-explanatory, and
wait for publication of the actual paper?
This sort of non-publication puts those of us who feel bound by ethical
considerations in an uncomfortable place.

I got several calls from the media trying to get a comment from me on this,
and they were spouting their usual pre-conceived narrative arcs of hype and
breathless sensationalism, so the informal announcement generated far more
heat than light, even though the intent of the study is to shed light on the
problem.

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