choice-metrics.com

Hi,

I am writing the code for the experimental design in Ngene. I am two numerical variables: sensitivity and frequency of screening (years). As I don't know whether these attributes will be continuous linear variables in the final model, is it more appropriate to code them as categorical (dummy coding) or continuous for the experimental design.

Thanks,
Amber

If you are not sure, then simply dummy code them as categorical attributes. This way, you will always be able to test for nonlinearities, while also still able to use it as a linear effect in model estimation instead.

Michiel