Re: looking for test of equivalence of covariances

On 19 Sep 2005 11:25:50 -0700, geraldmarschke@xxxxxxxxxxx wrote:

> I would like to test the equality of covariances (not covariance
> matrices) for two independent samples. The only tests I have found are
> for covariance matrices. Can anyone steer me in the right direction
> (offer references,software suggestions, etc)? Many thanks

I'll assume that you have the same two variables in two
samples -- I can't make sense of the question otherwise.

The test for "same regression coefficient" is the test that
seems reasonable. Again, I can't make much sense out
of a wider interpretation. My first thought was that you
should test the 2x2 covariance matrix, but you imply that
you don't want that.

To do this by regression, you put in a variable for Group,
a variable for Predictor, and a variable for interaction of
group-by-Predictor. The interaction is the test for Different
covariance, controlling for (potentially different) variances.

If you want something else, maybe you can be more
specific about the hypothesis -- variables and so on.

--
Rich Ulrich, wpilib@xxxxxxxx
http://www.pitt.edu/~wpilib/index.html
.

Relevant Pages