| Date: | Wed, 23 Jan 2008 11:49:40 -0800 |
|---|
| Reply-To: | "Earley, Joseph" <JEarley@LMU.EDU> |
|---|
| Sender: | "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU> |
|---|
| From: | "Earley, Joseph" <JEarley@LMU.EDU> |
|---|
| Subject: | Question on Existence of a test for the Significance of all
Regressors in systems of regression equations |
|---|
| Content-Type: | text/plain; charset="us-ascii" |
|---|
Hi All -
Is anyone familiar with whether or not a joint statistical test for the
statistical significance of each regressor in
each of say, five separate equations all tested at the same time exists
?
For example, say we have:
Y1 = bo + b1x1 + b2x2
Y2 = b00 + b3x3 + b4x4
.
.
.
Y5 = b + b9x9 + b10x10
Can we test if b1 through b10 are all significant at the same time.
I am familiar with the STEST option in PROC SYSLIN which tests the
equality of parameters in models fit to independent groups.
My question is slightly different.
Thanks in advance for any help.
Joe Earley
| 
