🚀 go-pugleaf

I'd appreciate advice on how to test this problem.  One item is drawn
from each of four populations.  Person 1 ranks the four items.  Person
2 also ranks the four items.  A new set of items is drawn.  Persons 1
and 2 again separately rank the four items.  In total, persons 1 and 2
rank 30 sets of the 4 items.  I would like to test whether persons 1
and 2 agree in their rankings.

If only one set of four items were used, it seems like Kendall's 
coefficient of concordance (W) would be appropriate.  I guess I could
average over the 30 sets of items for each person and then compute W.
But this doesn't seem to appropriately add power to the test to account
for 30 replications.  It also doesn't seem to control for each person
receiving the same sets of items. 

I'll post a summary.  Thanks in advance. 

--
Mike Kattan / Decision and Information Sciences / University of Houston 
Houston, TX  77204-6282 / (713) 743-4734 / fax 743-4693 / mkattan@uh.edu
==