Interpreting Stepwise result

[LaurenF]

Hi

I am busy with stepwise regression. I have already check that, individually, which variables are significant at a 5% level and then only included those variables in the Stepwise.

However on some of the equations, i am getting null values where in the Std Error, t-stat and prob columns and an R-squared of one.

Do you know why this could be happening? And if so, any tests that should be performed.

Many thanks
Lauren

[EViews Gareth]

Could you attach your workfile?

[LaurenF]

I think it may have something to do with the number of obs left for the stepwise, but I'm not sure. (see s_rev_cbh)

Also another problem I am having is that the stepwise regression does return a result (ie no variables are choosen, with r-squared of 0). However each variable on their own is statiscally significant (see s_capex_avi and s_capex_foodprod)

Your help is much appreciated.

[startz]

The problem is that you don't have enough valid observations in your sample.