Please excuse this cross posting--I think this thread runs from Arch-L to
Acra-L to Histarch. It is relevant to all, however.
 
 
In his book "Pioneer Pottery" Michael Cardew describes the principle of clay
sampling for potters:
 
"Sampling is defined as 'the art of extracting from a large bulk of material
a small portion which shall fairly represent the whole' (Mellor 1913).
Careless or inexpert sampling can nullify all the skill and care ogf the
analysts, since the analysis of a non-representative sample is worthless"
 
 
Ok We are archaeologists, not potters, but nonetheless there is food for
thought in this brief paragraph that has implications for us all.  A sample
must "fairly represent the whole"  or it will lead to worthless results,
regardless of the skills of the analyst.
 
Refer to Robert Keeler's Dissertation, or Stanley South's "Testing
Archaeological Sampling at Santa Elena" for some real world discussions of
sample representativeness.
 
Stopping now before I get carried away,
 
Carl Steen