Group dummy variables

[stefanp]

Hey there,

I have a bunch of dummy variables, say d_1_a, d_1_b,...,d_2_a, d_2_b etc. I would like to group them with something like:

Code: Select all

smpl @all for %j 1 2 3 4 for %k a b c d smpl @all string q = %j string r = %k group group_d_1_b "d_"+q+"_"+r next next

It doesnt really work, as the values of the group dummy variable are strings. I believe I mixed something up regarding the reference to the elements I want to group.

[EViews Gareth]

Can't really see what you're trying to do.

[stefanp]

I have several dummy variables which are named according to the following structure:

d_1_a
d_2_a
d_1_b
d_2_b

etc.

I would like to create a variable (e.g. group_dummy) that groups all but one of the dummy variables. The idea is that I don't want to include all dummy variables (i.e. not the group) in a regression equation.

Does that make more sense?

[EViews Glenn]

Code: Select all

group mygroup d_* mygroup.drop d_2_b equation eq01.ls y c d_2_b