- I can add something to what I have read in a couple of posts. On 3 Jun 2003 15:23:27 -0700, [EMAIL PROTECTED] (Carlos J. Vilalta y Perdomo) wrote:
> Would you please give me some advice in the following problem? > Problem: > Pearson correlation of x1 with y is n.s. > Yet, when y = x1 + x2 + x3... + x6 + e, then x1 becomes a significant > predictor of y > Questions: > 1. Would this be effect of an intervining/mediating variable? This *might* mean that the (x1,y) regression coefficient is exactly the same as the observed (x1,y) partial regression coefficient, where the latter equation has a (much?) smaller residual. For starters, the question lets us highlight why comparing "significant" to "non-significant" is an false basis of inference. Your example *might* describe a p-level of .0501 in the first case, and .0499 in the second; I hope you should play with enough data and equations to recognize how totally trivial that change would be. -- Rich Ulrich, [EMAIL PROTECTED] http://www.pitt.edu/~wpilib/index.html . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
