Thank you to everyone who responded to my stats query - it seems this is an
issue more contentious than I thought (and not as straightforward). A lot
of questions were posed, so perhaps some clarification on my part may help.
This is a random (as random as can be given the noted remarks on this)
sample of a pleistocene assemblage, with a total NISP of 1138 - and yes, it
is fragmented. Other evidence points to 2 distinct stratigraphic units,
formed under different conditions. Now this is where my problem just
begins. Not only are the formation processes different, but the larger
question is the role of human agency in the final assemblage. Given all the
other factors, my aim was to test these assumptions using element
frequencies. I did not know, however, if any of the usual tests would be
appropriate - given the use of NISP and the differences in formation
processes. Essentially, I was trying to find an "independent" test of these
assumptions using just the fauna.
I am aware of the merits of different quantification techniques, and
perhaps the upshot of this is that stats may have to wait for the moment
until MNE's can be calculated instead.
It is all too easy to argue over interpretations of an assemblage, so
quantifying the data w/ stats would seem to add strength to an arguement.
However, based on the discussions, I get the sinking feeling it may indeed
have the opposite effect.
Any and all other suggestions, are still greatly appreciated! Thank you to
everyone. Melanie
Dr. Melanie Fillios
University of Sydney
NSW 2006 Australia
[log in to unmask]
|