Displaying 1 result from an estimated 1 matches for "588e8bb9".
1999 Sep 09
1
princomp
Peter,
As I understand your Q. You probably have data that is similar to each other
like stock Prices for all RHS variable. In that case the difference between corr
and cov is not significant; however, if your RHS contains totally dissimilar
variables it matters a great deal. If x1 income, x2 job type, x3 Education
level, etc..., then taking cov of these variables would not be desireable