>>> Having considered all the answers received so far, I have concluded that
our  testing so far was  * sufficiently large for varroa <<<

>> Peter Edwards sent in a very nice table showing sample size and
certainty.
According to his table, with a sample size of 5 bees one could be 90%
certain *only* if the infestation rate was 50% or higher. I don't know how
you could conclude that a sample of 5 bees from the outside of the cluster
in winter was sufficient to *know* anything.<<

I can see you and I are looking at things from very different perspectives
and this is quite fascinating.  I am a commercial beekeeper.  Losing one
hive sixty miles away, especially if it is already over the hill, is of
academic interest to me only.

I think I covered some of this previously, but it probably bears repeating:
I am looking at the big picture and thinking of entire yards, rather than
individual hives.  In addition to the isolated results of this one test, I
have access to historical information, including recent treatments visual
assessments, and recent drop tests.  As a result, my required level of
certainty from this one test is considrably lower than if I had just
parachuted into strange terrain and had to start from scratch -- or try for
100% certainty and 100% survival.  I also have information about the
probable distribution of the pests in the hives, in the yards, and in the
operation and this affects how I analyse the data, and how I understand the
results of tests.

Perhaps it is time to reveal my own beliefs here: I regard this problem as a
multi-step problem, and not one where we can go direct to a definitive, 100%
correct answer.  We would love to have a simple litmus test, but that is not
in the cards.

I think this problem is somewhat recursive and that any solution is very
likely indeterminate, since there are factors and parameters we cannot hope
to predict or measure with the precison with which we can solve some easy
parts of the problem.  The math is easy, the assumptions are not.  We also
need to consider empirical data, but much of that appears condradictory.
The best we can hope for is to discuss and estimate ranges and thresholds,
and acceptable risk figures.

IMO, this type of decision making is often much more art than science. We
have to feed info to our neural networks and trust the results -- IMO.  What
I love about Bob's response (as did one detail-minded researcher who wrote
me privately after writing what he described as a 'tome') was that Bob was
not confused by the details or the analysis process.  He just scanned the
data given and his built-in 'Expert System' and Fuzzy Logic circuits said
"Probably OK".  He didn't give me a technical response, He just gave a
verdict and a few succinct reasons -- and caveats.

Maybe I like Bob's answer, too, because he agrees with me.  Is Bob right?
Am I right?  That is fun to analyze and very instructive, since that is how
we refine what we do and learn -- and teach.

That is one of the reasons that I thought it would be fun to put the
question to the list.  So we can all learn by banging heads.  If I
emphasized the statistical aspect of the problem, maybe that was a bit of an
unintentional red herring (as was pointed out by one astute bystander).  I
apologise, but I am fascinated by how much math we can apply to this problem
that is usually solved intuitively.  I like math.  Math works the same every
time.  For everyone, everywhere.  I doubt if simple math can give us a
definitive answer, but math can definitely solve parts of the problem
leaving less to intuition, and I have found the statistical contributions
very edifying.  I'm hoping for more.

> I still haven;t heard an explanation of why you would open the hives in
the middle of winter, or what you were expecting to find out... Am I missing
something?

I think I covered that in my (last?) post, but sometimes the LISTSERV
delivers the articles unevenly over time.  I know I have two accounts where
I receive articles and there is often a long span between when an article ap
pears in one and when it appears in the other.

allen
http://www.internode.net/honeybee/diary/