Subject: | |
From: | |
Reply To: | |
Date: | Tue, 5 Mar 2002 11:36:18 -0600 |
Content-Type: | text/plain |
Parts/Attachments: |
|
|
>Allen wrote
> >No guessing. It can be calculated.
I have been sitting thinking about this and I'm not sure you are right. I
don't think this can be calculated. To calculate sample size you need to
know first how the population is distributed, ie is it a normal / poisson /
exponential distribution? We would need to know our estimated and desired
test error rates and the confidence level we which to attain then plug
these into
n = (z^2 * p * q) / e^2
Where:
z = confidence limit factor (eg 1.96)
p = estimate of error rate (approximately 0.35)
q = 1 - p
e = desired error rate (eg +/- 5%)
However you really want 3 tests, varroa, nosema and tracheal mite.
Its doubtful that mites (spores) per bee is equally distributed in the
yard, we know one hive may have nosema and the others not.
We don't know what testing method you wish so we can't estimate the error
rates.
You have yards of hives instead of a single hive
We don't have total population in the hive
We don't know in any one of the 3 test the economic threshold, ie how many
mites ( spores) / bee is feasible to treat, what is a significant infection
level vs a nominal one.
Just thinking
Dave
|
|
|