logistic regression : intercept is NOT significative

[deleted_user]

Hi everybody

I have performed a logistic regression using a stepwise selection. But, at the end, the parameter estimate of intercept is NOT significative (p value of about 0.90).

What should I do ?

And how ?

Thanks for the help Message was edited by: babaorumi

[Paige]

Seriously, I wouldn't worry about it. Go ahead and use the model you came up with.

[Doc_Duke]

The intercept can be viewed as the baseline risk if all the other effects are set to their 0 value (reference cell coding for class variables). Whether it is meaningful or not depends on the sorts of independent variables and the domain of study. (It is usually not meaningful if there are any continuous predictors.)

I rarely even look at it.

Doc Muhlbaier
Duke