Does anyone know whether it is appropriate when using dummy variables in
regression analysis to deviate from a simple 1-0 coding system in order to
introduce a further element of analysis.
i.e. if we had categorical data set including FULL and EMPTY set at 1 or 0
would it be possible to introduce the concept of HALF-FULL alongside FULL.
(without setting HALF-FULL as a seperate variable)
or is this innapropriate or make the interpretation of the output too
difficult.
Thanks
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|