>>
Correlation problems are usually described if not defined as symmetric,
i.e. results are invariant to which variable is labelled as predictor
and which response; indeed qua correlation such labelling is
meaningless. And there's no question of an error term. Does discriminant
analysis match those details? >>
No, but cancor and discriminant analysis aren't obviously the same yet
they are mathematically. Obviously a single predictor generates only one
canonical correlation. I guess it depends on what you want to do. I
missed the original post (stupid me) so I probably expressed an opinion
on the wrong thing anyway.
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
