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?
Nick
[email protected]
Verkuilen, Jay
Tony Lachenbruch wrote:
>>When the discrete variable is a dichotomy, the test for the
correlation
is the same as the two-sample t-test with the dichotomy as the 'by'
variable. I've never looked at the case for discrete variables with k
categories. A little algebra could be fun and illuminating... (or
not)>>
Isn't this going to be a special case of discriminant analysis with one
continuous predictor?
*
* 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/
