st: Stepwise logistic regression with categorical variables

[Rachel Wade <Rachel.Wade@velindre-tr.wales.nhs.uk>]

Hello,

 

I am trying to carry out a stepwise logistic regression (I am aware of its limitations and am using it for exploratory purposes) and have a combination of categorical and continuous variables I wish to potentially include in a model. 

For example the line of code below:

xi: sw logistic status v3 i.v5 v6 v17 v18 v19 v21 i.v25 v26 i.v27 i.v28 i.v29 i.v30 i.v31

 

My question is how can I ensure that if the output contains a categorical variable (for example v25) that it contains the whole variable rather than just one or several of its categories, as this is no use to me.

 

I have looked in the help for xi and sw but find nothing on how to do this, yet I am sure that it must be possible.

 

Any assistance much appreciated,

 

With kind regards,

 

Rachel Wade