So this is what I need:
5 x 10ml samples (50ml total) from a single hive (a jar of honey is about 350ml for comparison).
Out of these 5 samples, 3 will be randomly selected for testing.
I'm assuming those 5 samples are from different locations in the hive, representing different times and sources of nectar?
If that is not the case, then wouldn't 3 samples be sufficient? Just trying to figure out the statistical advantage of randomizing samples from the same hive.
so typically, in studies, randomization is used to protect against selection bias...
Sampling from the same hive could be a pseudo-replicate and be meaningless to have so many samples. The original instructions just ask for 5 samples, which could include taking a big chunk of honey comb that gets split it into five samples. Randomizing the selection of 3 out of the 5 in this situation would be pointless. If the samples were from different frames, then there would be a difference in time and nectar source, which would make sense. I collected my samples from different frames just in case that is what he really wants. But it would be a lot easier to sample from the same frame, and only send 3 samples.
what you say makes sense, but as soon as one needs to publish a paper/thesis, the scientific establishment wants to see certain things... like how was the study designed wrt to randomization etc...
I understand the importance of randomization, but sampling from the same thing... is the same thing, and there is no possibility of bias because there is no difference between any of the samples. Publishing a paper is pretty strict, and methods need to be clear, like making sure replicates are truly replicates, and randomization is used appropriately. I occasionally am asked to review papers, (most recently the Journal of Economic Entomology) and this would be one of my first questions.
What I am really asking is for clarification on the methods of collection of the samples. I think he wants to have 5 samples from different locations from the hive and not 5 samples from the same frame.
If he doesn't need the samples from different locations, then it would be much easier to pop open the lid and take out 1 frame and get all the samples without hardly disturbing the hive. And the two extra samples would be unnecessary...saving time and postage expenses.
Collecting from different locations, requires a much more invasive pillaging of the hive. Taking samples from the different suppers... they are heavy right now, so it is a lot more work and gets the bees more rilled up. This would also give better data as to the situation of the hive. Still, three samples would be adequate to go through a peer review.
Just trying to get out of extra work or learn a new statistical lesson at the same time.
