Hello,
I'm posting a message to the mailbase out of desperation, not laziness
If I can briefly explain my problem regarding failure of independence:
I have a data set comprising a number of clonal organisms. These organisms
are made up of a number of units (usually between 1 and 10) of different
types (there are six common types, made up mostly of two of those types.
So for example, an organism could be made up of 3 types 2’s a type 4 and
two type 6’s).
What is of interest here is an apparent size difference between the
different types of unit, relative to the size of the other units in the
same organism.
These units are physiologically connected, therefore exhibit a failure of
independence, which presents a problem with determination of whether there
are differences in the relative sizes of the unit types.
So far I’ve been approaching the problem as follows:
By using a two stage sampling regime (randomly sampling one unit from each
randomly chosen organism, for a given percentage of the organisms in the
population, without replacement) I can obtain a sample of independent
units, which can be divided into the different types of units. ANOVA on
this sample reveals that there are significant differences between the
relative sizes of different unit types.
But this sample, although comprising independent units, is only a very
small fraction of the population, and n for several of the unit types is
very low in the sample. If the population were re-sampled, the units
comprising the sample would be independent from all those in the sample,
but not independent from those in the first sample.
So I’m stuck at this point.
I’d be really grateful for any suggestions on how to test whether there are
significant differences between in the relative size of the different unit
types, without making a violation of assumptions too far.
Thank you
David Oatway
----------------------------------------------------------------------------
David Oatway
Centre for Land Use and Water Resources Research
Porter Building
University of Newcastle
Newcastle upon Tyne
NE1 7RU
United Kingdom
Tel: +44 191 222 5956
Fax: +44 191 222 6563
URL: http://www.cluwrr.ncl.ac.uk
----------------------------------------------------------------------------
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|