Page 1 of 1

StepWise Regression

Posted: Wed Aug 03, 2011 1:07 pm
by jennie
Hi,

I am trying to run a stepwise regression:

equation eq.stepls(method=stepwise, back, btol=0.10, ftol=0.10) depName c @ sreit na01 pa gender age courses

Where
depName is the dependent variable for the regression
c is the constant that I would like to always have included
sreit na01 pa gender age courses is the list of independent variables that I am searching through

When I run the above equation I get the results

c 0.5038 0.0002
na01 -0.3548 0.3487
gender 11.9712 0.009
courses 1.1210 0.0599

(I have attached a copy of the results)

Why does the stepwise regression retain the na01 variable?
What I am trying to do is run the stepwise regression backwards and only keep p-values that are less than 0.10.

Is there something incorrect in my options specifiation? (method=stepwise, back, btol=0.10, ftol=0.10)
Or am I misunderstanding how the stepwise least squares works?

Please help ,

Jen

Re: StepWise Regression

Posted: Wed Aug 03, 2011 1:33 pm
by EViews Gareth
Probably a sample issue. The final regression output doesn't use the same sample as was used during the selection process. See the User Guide II for more details.