David C. Howell <[EMAIL PROTECTED]> wrote: : On one of my web pages I describe an algoritm for generating data with a : specified correlation between X and Y. The important part of the algorithm is
The general algorithm for a correlation matrix R is a trivial matrix exercise If a set of variables are uncorrelated with sd = 1, the vcvar matrix is I. The vcvar matrix of Ax is AA'. For any matrix R = AA', A is the required matrix and is most easily obtained as a Cholesky factor. If x is random, Ax gives random variables from a population with vcvar = R . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
