Recall that since my first publications (1986, 1991) on CR I said the assumptions of CR are::
1.Uniformly sampled causes (i.e. to form factorial designs). 2 Causes are uncorrelated (i.e. factorial form). 3 Effects correlated between .95 and .30 with causes. 4. Effects measured on continuous scale. I went back and looked at the data that Steve and Gottfried used to test CR. In every case, the causes are far from uniform and can not form factorial designs. EVERY CASE. Each of the tests... those that support and those that fail to support CR, are thus INVALID. I will bet David Heiser does not have the integrity to point this out to semnet folks. Dr. Bill . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
