REML ... JMP6 v. JMP7

[Robert Kinley <KINLEY_ROBERT@LILLY.COM>]

hi

Using the same script on the same data,
fitting a mixed model and
using REML, I get different results
depending whether I use version 
6 or version 7.

The log for version 6 (only) has some
comments about missing 
combinations of treatments, so I'm guessing
that the REML 
algorithm was improved between version
6 & 7, although I 
can't find anything about it on the
JMP website.

Does anyone know anything about this
...

        thanks
               Bob
 


~~~~~~~~~~~~~~~~
Robert  Kinley
Senior Research Scientist
Early Phase Statistics
Erl Wood, UK
tel    +44 1276 483393
mob  +44 7793 305879
fax   +44 1276 483525
~~~~~~~~~~~~~~~~