From: Andre Couturier <couturie@icm.umontreal.ca>
To: Frank E Harrell Jr <fharrell@virginia.edu>
Subject: Re: datasets
Date: Friday, November 06, 1998 2:05 PM

Dear Frank,

    you might remember that I asked you a question about 1 value from the prostate cancer
dataset,
which was coded as 9999 (The AP value for subject 279). You suggested I ask Dr. Green,
which in turn suggested I ask Donald K. Corle, which provided the following answer:

  <<I'm sure that this is a missing (unknown) observation.
  Actually, I'm a little confused.  Our convention was
  always to code missing as '9999', '99' , etc depending
  the number of digits recorded for the data item.  I am
  not sure why you are seeing '-9999' ??>>

  You can confirm with him at the following address:  dc13a@nih.gov


Have a good weekend

Andre Couturier