[buug] Determining which records were dropped?

[Claude Rubinson]

---- Bob wrote ----
>The path of least work might be to take the dataset
>"as is" and recalculate.
>
>In other words, who cares what the other guy
>did...you've got the data now...it's YOUR turn to kick
>some ass.

Actually, it looks like that's what we're going to do. 
Apparently, the previous analyst was known to be a bit sloppy.
 So people seem to be willing to take my results over his. 
(I'd still like to know what he did, however.  But I'm pretty
Type-A.)