[EMAIL PROTECTED] (Herman Rubin) wrote in message
> Without more assumptions, the question cannot be answered. > > One can give probability models in which it can be anything. What sort of assumptions would need to be made, what kind of probability models are your referring to? . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
