Hello,
I have been asked to send this question to the list by a colleague:
We have a large geographical area split into 4 regions, A,B,C and D.
These are further divided into a number of sections:
a_1,a_2,a_3...a_j
b_1,b_2...b_k
c_1...c_m
d_1...d_n
Finally, these are further subdivided into a number of units:
a_1_1, a_1_2... And so on.
Each of these units has a measure of success associated to it.
Furthermore, each of the sections and regions have similarly calculated
measures of success associated with them.
The area as a whole is thought to be probably non-homogenous with respect to
a number of other factors. These factors would vary over the whole area,
but are not necessarily entirely predictable along the lines of regions and
sections and are not necessarily similar within regions or sections.
My friend wishes to try and determine how much variation between regions is
due to variation between sections and units, and how much is due to "genuine"
inter region variation. I dont think I've managed to express my friends
question
very well, and I would like to apologise for that, but I would appreciate any
help that anyone felt able to give.
Thanks,
Jill
|