On Sun, 2 Jan 2000, Christopher wrote:
> For 2 continuous variables, we measure the linear relation through
> correlation. For 2 categorical variables, we measure the relation
> through chi-square test.
> But how to measure, non-linear relation for 2 continuous variables.
Two general approaches.
(1) If you know the functional relation Y = f(X) + error
you can model it via multiple regression analysis, which
will report R-sq, the square of the correlation between Y
and f(X).
(If you don't _know_ the functional relation, but have some
intelligent guess(es) as to what it might be, you can model
the function(s) you think reasonable, and pick one, or a few,
for further study.)
(2) If you're altogether at sea about possible functions, you can
stratify X into a number of more or less homogeneous slices
(your choice of how many), use the slices as levels of a
categorical variable for a 1-way analysis of variance, from
which you can compute either eta-squared or the equivalent of
R-square (= SSbetween/SStotal).
But if the relation between X and Y is such that Y cannot be
expressed as a function of X, the problem is less easy to describe in
general terms.
-- DFB.
------------------------------------------------------------------------
Donald F. Burrill [EMAIL PROTECTED]
348 Hyde Hall, Plymouth State College, [EMAIL PROTECTED]
MSC #29, Plymouth, NH 03264 603-535-2597
184 Nashua Road, Bedford, NH 03110 603-471-7128
