Regression Coefficient Standard Error
 

On Tuesday, February 11, 2003 11:45:34 PM UTC-5, Jerry W. Lewis wrote:
Agreed.

Harlan Grove wrote:

"Jerry W. Lewis" wrote...

http://groups.google.com/groups?selm...40mediaone.net

provides formulas that do not require array formulas and are more
numerically accurate than the equations Harlan suggested. His approach
is mathematically exact, but numerically unstable.


Precisely! The OP asked for what Excel uses, and that's what I gave.
Mathematically exact but numerically unstable, just like Excel.

--
Public Service Announcement
Don't attach files to postings in nonbinary newsgroups like this one.

If you have multiple regressors the SLOPE and INTERCEPT functions will not work. I can't see how to avoid Matrix algebra in that case.