On Jan 18, 2004, at 2:33 AM, Richard wrote:
Now, I've read that there are different ways of computing the eigenvalues
and condition indices; can anybody offer any insights as to what should
generally be preferred or when you should prefer one approach over
another?
Read the Belsley and Welsch book (or the earlier Belsley, Welsch and Kuh)
cited in the Stata manual (and in most textbooks in their discussion of
collinearity issues).
Thanks Kit. I found another Stata module, -coldiag-, which gives yet a
third estimate of the condition number! But it, too, gives the same
results as the other methods when variables are centered first.
