[buug] Determining which records were dropped?

Claude Rubinson cmsclaud at uga.edu
Wed Jul 2 11:46:06 PDT 2003


---- Jon wrote ----
>Do you know _why_ eight records were dropped? That would get 
>you a long ways towards figuring out which 8 to drop. Have you 
>tried looking at the 58 records to see if there's anything in 
>common amongst a subset? You could also just try common
>things, like dropping outliers, etc.

That's the problem -- I've tried everything obvious without
success.  And there are 245 variables per record -- too many
criteria permutations to be of use.  (Frankly, my growing fear
is that something else was done to the dataset which I'm not
aware of.  I.e., even if I do hit on the correct subset, I
still won't get the expected results.  But at this point, I
can't be sure of that -- it would at least be nice to be able
to say definitively that the dataset has been changed.)



More information about the buug mailing list