Date: Dec 7, 2012 9:02 AM
Author: Halitsky
Subject: Please ignore my first question about "estimated standard errpr" in<br> my last post !!!! Sorry !

I misread my own code due to a stupid MS Word line spacing error when
I looked at it in Windows rather than Linux.

The formula that you gave me for the 2-way t has the sqrt of the sum
of the variances in the denominator:

(coeff1-coeff2) + (coeff3-coeff4)
t = --------------------------------------------
SQRT(var1 + var2 + var3 + var4)

So it makes sense now that in the single coefficient formula you just
provided, the denominator is simply the SE of the coefficient.