[EMAIL PROTECTED] (Garbriel) wrote in news:[EMAIL PROTECTED]:
> Thanks for your help, > You mean there is no approximation method, like adding 0.5 to each > cell, to do this case? Nope. > If there is no varation in one varible. I want know that if there is > relation between these 2 variable, How can I do? You simply can't, any more than you can sleep awake or sit standing up. If one of your variables doesn't vary, you can't talk about how its variation corresponds to variation in the other. If this is an inferential problem, you simply don't have enough information to draw *any* conclusions. If it's a descriptive problem, all you can say is that one of your variables is actually a constant and therefore doesn't convey any information. In the latter case, your best bet is to demonstrate statistical independence (a stronger condition than correlation). . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
